Reproduce “PDE-OBS: Controlled Evaluation Across Observation Patterns” 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
Ruichen Xu, Siyao Wang, Fang Wan, Jiacheng Qiu, Wenhan Gao, Jiaxing Zhang, Linsey Pang, Ravid Shwartz-Ziv and 2 more
PDE-OBS tests whether physical-field reconstruction and forecasting models transfer across observation patterns. It keeps records, complete targets, and splits fixed while changing binary masks, including random, grid, line, block, and clustered layouts. Across 49 PDE–method pairs, cross-pattern error exceeded matched-pattern error, with median ratio 10.46. This shows sensitivity in a selected, noiseless, smooth-field setup, not deployment performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.