Data-driven discovery of Green’s functions with human-understandable deep learning

• Main Authors: Boullé, N. (Author), Earls, C.J (Author), Townsend, A. (Author)
• Format: Article
• Language: English
• Published: Nature Research 2022
• Subjects: advection; article; deep learning; Deep Learning; diffusion; excitation; human; human experiment; Humans; machine learning; Machine Learning; Neural Networks, Computer; normal distribution; Normal Distribution; physics
• Online Access: View Fulltext in Publisher

 There is an opportunity for deep learning to revolutionize science and technology by revealing its findings in a human interpretable manner. To do this, we develop a novel data-driven approach for creating a human–machine partnership to accelerate scientific discovery. By collecting physical system responses under excitations drawn from a Gaussian process, we train rational neural networks to learn Green’s functions of hidden linear partial differential equations. These functions reveal human-understandable properties and features, such as linear conservation laws and symmetries, along with shock and singularity locations, boundary effects, and dominant modes. We illustrate the technique on several examples and capture a range of physics, including advection–diffusion, viscous shocks, and Stokes flow in a lid-driven cavity. © 2022, The Author(s).