Reproduce “Physics-Informed Neural Plasticity: PDE Solvers That Reshape Themselves” 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
Chun-Wun Cheng, Bingcheng Hu, Angelica I. Aviles-Rivero
ReCAP adapts local Gaussian components with polynomial representations during PINN training. Residuals guide refinement, splitting, pruning, and merging, while transported coefficients let new children initially preserve a parent’s local polynomial. Across five tested three- and four-dimensional PDE benchmarks, ReCAP reports the lowest relative L2 error against each task’s strongest competitor; on KG3D, it reports 0.003181, versus SPINN’s 0.003773.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.