Reproduce “Geometry-physics confounding impairs PDE learning across varying domains” 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
Yinghao Cheng, Gengxiang Chen, Xu Liu, Qinglu Meng, Yixin Jing, Xiangguo Tang, Wenping Mou, Lihui Wang and 1 more
Geometry–physics confounding affects PDE learning across changing domains. Authors provide map-induced coefficient fields for operator learning and transformed candidate operators for sparse equation discovery. Geometry-conditioned models had lower test error at every tested training size across five benchmarks. In airfoil flow, error was 36.5% lower; in tissue discovery, residual was over two orders lower, using known geometry and boundary-motion closure.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.