ORCID

0000-0003-4162-0276 (Colen)

Document Type

Article

Publication Date

2024

DOI

10.1103/PhysRevLett.133.107301

Publication Title

Physical Review Letters

Volume

133

Issue

10

Pages

107301 (1-7)

Abstract

We present a data-driven pipeline for model building that combines interpretable machine learning, hydrodynamic theories, and microscopic models. The goal is to uncover the underlying processes governing nonlinear dynamics experiments. We exemplify our method with data from microfluidic experiments where crystals of streaming droplets support the propagation of nonlinear waves absent in passive crystals. By combining physics-inspired neural networks, known as neural operators, with symbolic regression tools, we generate the solution, as well as the mathematical form, of a nonlinear dynamical system that accurately models the experimental data. Finally, we interpret this continuum model from fundamental physics principles. Informed by machine learning, we coarse grain a microscopic model of interacting droplets and discover that non-reciprocal hydrodynamic interactions stabilise and promote nonlinear wave propagation.

Rights

© 2024 American Physical Society.

The author or the author's employer may use all or part of the APS published article, including the APS-prepared version (e.g., the PDF from the online journal) without revision or modification, on the author's or employer's website as long as a fee is not charged. If a fee is charged, then APS permission must be sought. In all cases, the appropriate bibliographic citation and notice of the APS copyright must be include.

Original Publication Citation

Colen, J., Poncet, A., Bartolo, D., & Vitelli, V. (2024). Interpreting neural operators: How nonlinear waves propagate in nonreciprocal solids. Physical Review Letters, 133(10), Article 07301. https://doi.org/10.1103/PhysRevLett.133.107301

Additional Files
supplement.pdf (1317 kB)
Download

Share

 
COinS