An AI physics model for near-real-time, high-fidelity aerodynamic simulations from raw geometry inputs without CFD re-meshing.

A unified deep learning framework for compressing and simulating diverse spatio-temporal physics problems across grid and particle methods.