Reproduce “Feature tracking in physics-informed neural networks via joint optimization of nonlinear deformation manifolds: application to shocks” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Akshay Thakur, Matthew Zahr
FT-PINN learns a smooth, non-separable deformation of a fixed grid alongside a neural solution, using conservation-law residuals to move points toward features without presetting their locations. In a steady two-dimensional Euler shock-reflection test, it recovers both shocks with 5,684 interior points, where a matched-grid vanilla PINN forms neither; a 12,000-point random baseline forms broad shocks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.