Neural Physical Simulation with Multi-Resolution Hash Grid Encoding
DOI:
https://doi.org/10.1609/aaai.v38i6.28349Keywords:
CV: 3D Computer Vision, APP: Natural Sciences, CSO: Solvers and ToolsAbstract
We explore the generalization of the implicit representation in the physical simulation task. Traditional time-dependent partial differential equations (PDEs) solvers for physical simulation often adopt the grid or mesh for spatial discretization, which is memory-consuming for high resolution and lack of adaptivity. Many implicit representations like local extreme machine or Siren are proposed but they are still too compact to suffer from limited accuracy in handling local details and a long time of convergence. We contribute a neural simulation framework based on multi-resolution hash grid representation to introduce hierarchical consideration of global and local information, simultaneously. Furthermore, we propose two key strategies: 1) a numerical gradient method for computing high-order derivatives with boundary conditions; 2) a range analysis sample method for fast neural geometry boundary sampling with dynamic topologies. Our method shows much higher accuracy and strong flexibility for various simulation problems: e.g., large elastic deformations, complex fluid dynamics, and multi-scale phenomena which remain challenging for existing neural physical solvers.Downloads
Published
2024-03-24
How to Cite
Wang, H., Yu, T., Yang, T., Qiao, H., & Dai, Q. (2024). Neural Physical Simulation with Multi-Resolution Hash Grid Encoding. Proceedings of the AAAI Conference on Artificial Intelligence, 38(6), 5410-5418. https://doi.org/10.1609/aaai.v38i6.28349
Issue
Section
AAAI Technical Track on Computer Vision V