Machine learning for ultrafast photonics applications: from nonlinear instabilities to broadband supercontinuum generation
We review the use of machine learning techniques in ultrafast photonics applications with emphasis on fiber-optics systems. In particular, we discuss how neural networks can be used to extract quantitative time-domain information in the development of nonlinear instabilities from spectral intensity measurements. We also show how neural networks can be efficiently applied to predict nonlinear dynamics in optical fibres for a wide range of scenarios, from pulse compression to ultra-broadband supercontinuum generation in both single and multimode fibers.
Tampere Univ. (Finland)