Reproduce “Hybrid Lagrangian-Eulerian Model for Lagrangian Fluid Simulation” 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
Ruoyan Li, Wei Wang, Yizhou Sun
The authors propose a hybrid Lagrangian–Eulerian neural simulator for fluids with moving boundaries. It keeps particles, aggregates their features onto fixed Eulerian nodes, and combines self- and cross-attention before decoding particle predictions. On Water, at 200 steps, kinetic-energy error is 3.735 versus 1.545 × 10³ for GraphUnet. This comparison used auxiliary training strategies and does not isolate cross-attention.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.