#adaptive time-stepping

Texas A&M researchers developed ShockCast, a two-phase machine learning approach that adaptively predicts timesteps to improve high-speed fluid flow simulations with neural networks.