no code implementations • 21 Nov 2022 • Chloe Paliard, Nils Thuerey, Kiwon Um
We explore training deep neural network models in conjunction with physics simulations via partial differential equations (PDEs), using the simulated degrees of freedom as latent space for a neural network.
no code implementations • 29 Sep 2021 • Chloé Paliard, Nils Thuerey, Marco Cagnazzo, Kiwon Um
In contrast to previous work, we do not constrain the PDE solver but instead give the neural network complete freedom to shape the PDE solutions as degrees of freedom of a latent space.
4 code implementations • 11 Sep 2021 • Nils Thuerey, Philipp Holl, Maximilian Mueller, Patrick Schnell, Felix Trost, Kiwon Um
This digital book contains a practical and comprehensive introduction of everything related to deep learning in the context of physical simulations.
no code implementations • 20 Nov 2020 • Marie-Lena Eckert, Kiwon Um, Nils Thuerey
In this paper, we present ScalarFlow, a first large-scale data set of reconstructions of real-world smoke plumes.
3 code implementations • NeurIPS 2020 • Kiwon Um, Robert Brand, Yun, Fei, Philipp Holl, Nils Thuerey
Finding accurate solutions to partial differential equations (PDEs) is a crucial task in all scientific and engineering disciplines.
1 code implementation • ICML 2020 • Georg Kohl, Kiwon Um, Nils Thuerey
We propose a neural network-based approach that computes a stable and generalizing metric (LSiM) to compare data from a variety of numerical simulation sources.
no code implementations • ICLR 2020 • Kiwon Um, Yun (Raymond) Fei, Philipp Holl, Nils Thuerey
While our approach is very general and applicable to arbitrary partial differential equation models, we specifically highlight gains in accuracy for fluid flow simulations.
1 code implementation • 14 Apr 2017 • Kiwon Um, Xiangyu Hu, Nils Thuerey
We use neural networks to model the regression of splash formation using a classifier together with a velocity modifier.