Figure 1
From: Machine learning analysis of rogue solitons in supercontinuum generation
(a) Schematic of machine learning application to the analysis and prediction of rogue soliton temporal characteristics in supercontinuum generation based only on single-shot spectral measurements. (b–e) Dynamics of noisy supercontinuum generation. Single-shot spectral (b) and temporal (c) evolution where we see enhanced spectral broadening associated with the emergence of a rogue soliton. (d,e) show a selection of 5 random supercontinuum spectra and corresponding temporal intensity profiles in black. The blue curve in (e) shows the rogue soliton temporal intensity associated with the long-pass filtered spectral components (blue area) in (d) while the black dotted line shows the full time-domain intensity.
