Inverse design of a Raman amplifier in frequency and distance domains using convolutional neural networks

Abstract

We present a Convolutional Neural Network (CNN) architecture for inverse Raman amplifier design. This model aims at finding the pump powers and wavelengths required for a target signal power evolution, both in distance along the fiber and in frequency. Using the proposed framework, the prediction of the pump configuration required to achieve a target power profile is demonstrated numerically with high accuracy in Cband considering both counter-propagating and bidirectional pumping schemes. For a distributed Raman amplifier based on a 100 km single-mode fiber, a low mean set (0.51, 0.54 and 0.64 dB) and standard deviation set (0.62, 0.43 and 0.38 dB) of the maximum test error are obtained numerically employing 2 and 3 counter, and 4 bidirectional propagating pumps, respectively.

Modeling optical amplifiers: from inverse design to full system optimization

“…Finally, inverse modeling of Raman amplifiers can be extended to focus on the amplifier response over both frequency and distance. This has been proposed [11] and experimentally validated [12], to provide additional flexibility in the optimization of distributed Raman amplification over transmission links, e.g. for applications such as nonlinearity compensation techniques [13].…”

A Framework for Biosensors Assisted by Multiphoton Effects and Machine Learning

“…This challenging optimization issue calls for the solution of a set of nonlinear differential equations. Many algorithms have been developed [ 224 ], as well as ANN [ 225 ] or ML [ 226 ] to find a solution to the conflict between the pump setting and the intended spectral gain setting. Currently, an ML strategy has been proposed for single-mode fibers [ 227 ] and few-mode fibers [ 225 ].…”

“…Many algorithms have been developed [ 224 ], as well as ANN [ 225 ] or ML [ 226 ] to find a solution to the conflict between the pump setting and the intended spectral gain setting. Currently, an ML strategy has been proposed for single-mode fibers [ 227 ] and few-mode fibers [ 225 ]. A dataset of hundreds of advantage bends made with erratic pump powers and wavelengths is used to train an NN to consider the relationship between the pump parameters [ 228 ].…”

Flexible Raman Amplifier Optimization Based on Machine Learning-Aided Physical Stimulated Raman Scattering Model

“…State of the art approaches for optimization are also based on heuristics [4], [9], [10]. Machine learning (ML) methods for the optimization are also gaining traction [11], [12], [13], [14], especially methods which train combinations of forward and inverse system models to predict the required pump power and frequency for a given target gain profile [11], [15], [16], [17], [18]. ML methods provide excellent performance, however, they require a lot of training data to be generated in order to populate the 2N p dimensional space, where N p is the number of pumps.…”