Automating physical intuition in nonlinear fiber optics with unsupervised dominant balance search

TL;DR

This paper presents an unsupervised machine learning method to automatically identify dominant physical interactions in nonlinear fiber optics, enhancing understanding of complex optical phenomena like soliton fission and wavebreaking.

Contribution

It introduces a novel unsupervised dominant balance search technique for analyzing nonlinear optical dynamics, providing automated insights into complex physical processes.

Findings

Successfully identified dominant interactions in various nonlinear optical phenomena

Demonstrated the method's ability to analyze temporal and spectral domains

Enhanced understanding of physical mechanisms in fiber optics

Abstract

Identifying the underlying processes that locally dominate physical interactions is the key to understanding nonlinear dynamics. Machine-learning techniques have recently been shown to be highly promising in automating the search for dominant physics, adding important insights that complement analytical methods and empirical intuition. Here we apply a fully unsupervised approach to the search for dominant balance during nonlinear and dispersive propagation in optical fiber, and show that we can algorithmically identify dominant interactions in cases of optical wavebreaking, soliton fission, dispersive wave generation, and Raman soliton emergence. We discuss how dominant balance manifests both in the temporal and spectral domains as a function of propagation distance.