Interpreting neural operators: how nonlinear waves propagate in non-reciprocal solids

We introduced a modeling pipeline that combines interpretable machine learning and continuum theory to guide discovery of new physics. Using a microfluidics experiment as a case study, we learned nonlinear dynamics directly from data using neural operators and sparse regression. We extracted the governing equations and showed how they originated from non-reciprocal hydrodynamic interactions between droplets.

Collaboration with Alexis Poncet (ENS Lyon), Denis Bartolo (ENS Lyon), and Vincenzo Vitelli (UChicago).