We propose a novel machine learning based approach that formulates physics-based fluid simulation as a regression problem, estimating the acceleration of every particle for each frame. Our GPU implementation runs in real-time for systems with up to 2 million particles.

Lubor Ladicky, ETH Zurich
SoHyeon Jeong, ETH Zurich
Barbara Solenthaler, ETH Zurich
Marc Pollefeys, ETH Zurich
Markus Gross, Disney Research, ETH Zurich