Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group
Journal Home About Issues in Progress Current Issue All Issues Early Posting

Deep neural network for modeling soliton dynamics in the mode-locked laser

Yin Fang, Hao-Bin Han, Wen-Bo Bo, Wei Liu, Ben-Hai Wang, Yue-Yue Wang, and Chao-Qing Dai

Open Access Open Access

Abstract

Integrating the information of the first cycle of an optical pulse in a cavity into the input of a neural network, a bidirectional long short-term memory (Bi_LSTM) recurrent neural network (RNN) with an attention mechanism is proposed to predict the dynamics of a soliton from the detuning steady state to the stable mode-locked state. The training and testing are based on two typical nonlinear dynamics: the conventional soliton evolution from various saturation energies and soliton molecule evolution under different group velocity dispersion coefficients of optical fibers. In both cases, the root mean square error (RMSE) for 80% of the test samples is below 15%. In addition, the width of the conventional soliton pulse and the pulse interval of the soliton molecule predicted by the neural network are consistent with the experimental results. These results provide a new insight into the nonlinear dynamics modeling of the ultrafast fiber laser.

© 2023 Optica Publishing Group

Full Article  |  PDF Article
More Like This
Design and analysis of recurrent neural networks for ultrafast optical pulse nonlinear propagation

Gustavo R. Martins, Luís C. B. Silva, Marcelo E. V. Segatto, Helder R. O. Rocha, and Carlos E. S. Castellani
Opt. Lett. 47(21) 5489-5492 (2022)

Mutual dynamics between synchronous solitons in a bidirectional mode-locked fiber laser

Yujia Li, Chao Wang, Dongmei Huang, Hongjie Chen, and Feng Li
Opt. Lett. 47(9) 2170-2173 (2022)

Unveiling external motion dynamics of solitons in passively mode-locked fiber lasers

Yusheng Zhang, Lin Huang, Yudong Cui, and Xueming Liu
Opt. Lett. 45(17) 4835-4838 (2020)

View More...

Supplementary Material (2)

NameDescription
Supplement 1       Supplemental Document
Visualization 1       Supplemental Document

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (5)
Fig. 1.
Fig. 1. Flow chart of the DNN to predict soliton dynamics. (a) GNLSE simulation studies the transient variations in the ultrafast laser; (b) schematic diagram of the Bi_LSTM RNN.

Download Full Size | PDF

Fig. 2.
Fig. 2. Conventional soliton predicted by the RNN with the saturation energy ${E_{sat}} = 16.5\textrm{ pJ}$. (a) Change of energy along the propagation distance with the insets showing the dynamics prediction; (b) width evolution of pulse via the GNLSE simulation and RNN prediction.

Download Full Size | PDF

Fig. 3.
Fig. 3. Conventional soliton predicted by the RNN with the saturation energy ${E_{sat}} = 17.8\textrm{ pJ}$. (a) 1000-cycle dynamics of conventional soliton predicted by the RNN; (b) energy, and (d) width evolution of pulse via the GNLSE simulation and RNN prediction; (c) comparison of pulse width between the experimental autocorrelation data and RNN prediction.

Download Full Size | PDF

Fig. 4.
Fig. 4. Soliton molecule predicted by the RNN and measured in the experiment. (a) Pulse evolution, and (b) autocorrelation trajectory predicted by the RNN; (c) experimental autocorrelation graph; (d) autocorrelation graph in the last lap predicted by the RNN.

Download Full Size | PDF

Fig. 5.
Fig. 5. Soliton molecule. (a) Evolutions of two solitons via the GNLSE simulation and RNN prediction. (b) RMSE of samples predicted by the RNN, under the group velocity dispersion coefficient ${\beta _2} = 13.782 \times {10^{ - 5}}\textrm{ p}{\textrm{s}^\textrm{2}} \cdot \textrm{k}{\textrm{m}^{\textrm{ - 1}}}$. (c) Dynamics, and (d) three-dimensional error of soliton molecule from 450 to 1500 cycles predicted by the RNN under the group velocity dispersion coefficient ${\beta _2} = 13.754 \times {10^{ - 5}}\textrm{ p}{\textrm{s}^\textrm{2}} \cdot \textrm{k}{\textrm{m}^{\textrm{ - 1}}}$.

Download Full Size | PDF

Tables (1)

Tables Icon

Table 1. Comparison of CPU and RAM Resources during the GNLSE Simulations and RNN Training

View Table

Equations (1)

Equations on this page are rendered with MathJax. Learn more.

R M S E = ∑ c , l ( x m , c , l − x ^ m , c , l ) 2 ∑ c , l ( x m , c , l ) 2 ,
Select as filters
Select Topics Cancel
Top
       
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.
Privacy | Terms of Use