Reusability report: Predicting spatiotemporal nonlinear dynamics in multimode fibre optics with a recurrent neural network

Teğin, Uğur; Dinç, Niyazi Ulaş; Moser, Christophe

doi:10.1038/s42256-021-00347-6

Cited by 25 publications

(11 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The triple multiplications (AlAmAn*) with varying strengths of interaction (ηp,l,m,n) between different sets of modes demonstrate the nonlinear and immensely multimodal aspect of the interactions such that modeling the transform of a single pulse on the setup provided in Fig. 1 would take 50 min with a graphics processing unit (GPU) 42 . The physical optical system carries out this complex spatiotemporal transformation “effortlessly.” The transformation is programmed with a relatively small number of PPs.…”

Section: Methods and Resultsmentioning

confidence: 99%

Programming nonlinear propagation for efficient optical learning machines

Oguz,

Hsieh,

Dinc

et al. 2024

Adv. Photon.

View full text Add to dashboard Cite

.The ever-increasing demand for training and inferring with larger machine-learning models requires more efficient hardware solutions due to limitations such as power dissipation and scalability. Optics is a promising contender for providing lower power computation, since light propagation through a nonabsorbing medium is a lossless operation. However, to carry out useful and efficient computations with light, generating and controlling nonlinearity optically is a necessity that is still elusive. Multimode fibers (MMFs) have been shown that they can provide nonlinear effects with microwatts of average power while maintaining parallelism and low loss. We propose an optical neural network architecture that performs nonlinear optical computation by controlling the propagation of ultrashort pulses in MMF by wavefront shaping. With a surrogate model, optimal sets of parameters are found to program this optical computer for different tasks with minimal utilization of an electronic computer. We show a remarkable decrease of 97% in the number of model parameters, which leads to an overall 99% digital operation reduction compared to an equivalently performing digital neural network. We further demonstrate that a fully optical implementation can also be performed with competitive accuracies.

show abstract

Section: Methods and Resultsmentioning

confidence: 99%

Programming nonlinear propagation for efficient optical learning machines

Oguz,

Hsieh,

Dinc

et al. 2024

Adv. Photon.

View full text Add to dashboard Cite

show abstract

“…Numerical studies. We performed simulations with the time dependent beam propagation method 42,43 to understand the effect of pulse power to spatial distribution of the pulse. Since the output beam profile of the fiber laser is heavily dependent on the GRIN MMF section, which is the last section before the NPE output of the laser, we studied the pulse propagating in this parabolic-index fiber.…”

Section: Resultsmentioning

confidence: 99%

“…In time-dependent beam propagation method simulations, heavy multidimensional fast Fourier-transform calculations require long computation times thus we utilized GPU-parallelization in our code. In our simulations we performed symmetrized split-step Fourier method to compute field propagation 42,43 . In the simulation, pulses with a Gaussian temporal distribution centered at 1064 nm with an 8 ps duration (full-width at half-maximum) are numerically propagated for a fiber length of 2 m. The launched beam diameters (1/e 2 ) are set to Yb-doped gain fibers core size (10 μm) with 15 μm offset.…”

Section: ∂U ∂Zmentioning

confidence: 99%

Real-time observation of optical rogue waves in spatiotemporally mode-locked fiber lasers

Teğin

Wang

2023

Commun Phys

Self Cite

View full text Add to dashboard Cite

Fiber lasers offer tabletop nonlinear environments to mimic and study the complex dynamics of nature. Optical rogue waves, rarely occurring extreme intensity fluctuations, are one of the many subjects that can be investigated with a fiber laser cavity. Although oceanic rogue waves are a result of spatiotemporal dynamics, the single-mode nature of the fiber laser and the commonly used measurement techniques limit the optical rogue wave studies to only temporal dynamics. In this study, we overcome such limit to observe rogue wave real-time dynamics in spatiotemporally mode-locked fiber lasers by utilizing state-of-the-art compressed ultrafast photography technique. The multimode laser cavity exhibits long-tailed non-Gaussian distributions under relaxed cavity constraints. Single-shot spatiotemporal measurements of rogue events showed that, instead of noise bursts, the cavity produces clean pulses with high-quality beam profiles. Our results indicate that rogue events in spatiotemporally mode-locked fiber lasers undergo nonlinear spatial transformation due to a power-dependent consistent attractor.

show abstract

“…There is a strong interest in finding a datadriven solution through machine learning. In recent years, machine learning has shown power in predicting complex nonlinear evolution governed by NLSE [73][74][75]. PINNs guided with specific theories can also be an effective analytical tool to solve PDEs from incomplete models and limited data [76].…”

Section: Predictionmentioning

confidence: 99%

“…Later in 2021, they extended the method of ref. [73] (see Pulse Prediction of Nonlinear Dynamics) to predict spatiotemporal nonlinear propagation for an arbitrary number of modes in graded-index multimode fibers through a RNN [74] (Fig. 8).…”

Section: Spatiotemporal Nonlinearities Prediction and Controlmentioning

confidence: 99%

Fiber laser development enabled by machine learning: review and prospect

Jiang

et al. 2022

PhotoniX

View full text Add to dashboard Cite

In recent years, machine learning, especially various deep neural networks, as an emerging technique for data analysis and processing, has brought novel insights into the development of fiber lasers, in particular complex, dynamical, or disturbance-sensitive fiber laser systems. This paper highlights recent attractive research that adopted machine learning in the fiber laser field, including design and manipulation for on-demand laser output, prediction and control of nonlinear effects, reconstruction and evaluation of laser properties, as well as robust control for lasers and laser systems. We also comment on the challenges and potential future development.

show abstract

Reusability report: Predicting spatiotemporal nonlinear dynamics in multimode fibre optics with a recurrent neural network

Cited by 25 publications

References 21 publications

Programming nonlinear propagation for efficient optical learning machines

Programming nonlinear propagation for efficient optical learning machines

Real-time observation of optical rogue waves in spatiotemporally mode-locked fiber lasers

Fiber laser development enabled by machine learning: review and prospect

Contact Info

Product

Resources

About