Automating Physical Intuition in Nonlinear Fibre Optics Using Dominant Balance Search

Artificial intelligence is profoundly transforming research in optics, with new techniques for experimental data analysis and the use of advanced methods of laser optimisation. While these experimental breakthroughs are remarkable, the impact of machine learning extends far beyond the laboratory, and there are equally revolutionary possibilities to advance theory. For example, machine learning algorithms can “discover” new physical laws from data, and clustering methods can greatly improve our understanding of nonlinear dynamics. In other words, our interpretation of physical systems need not depend solely on human experience and intuition.