Accelerating Particle and Fluid Simulations with Differentiable Graph Networks for Solving Forward and Inverse Problems (2309.13348v1)
Abstract: We leverage physics-embedded differentiable graph network simulators (GNS) to accelerate particulate and fluid simulations to solve forward and inverse problems. GNS represents the domain as a graph with particles as nodes and learned interactions as edges. Compared to modeling global dynamics, GNS enables learning local interaction laws through edge messages, improving its generalization to new environments. GNS achieves over 165x speedup for granular flow prediction compared to parallel CPU numerical simulations. We propose a novel hybrid GNS/Material Point Method (MPM) to accelerate forward simulations by minimizing error on a pure surrogate model by interleaving MPM in GNS rollouts to satisfy conservation laws and minimize errors achieving 24x speedup compared to pure numerical simulations. The differentiable GNS enables solving inverse problems through automatic differentiation, identifying material parameters that result in target runout distances. We demonstrate the ability of GNS to solve inverse problems by iteratively updating the friction angle (a material property) by computing the gradient of a loss function based on the final and target runouts, thereby identifying the friction angle that best matches the observed runout. The physics-embedded and differentiable simulators open an exciting new paradigm for AI-accelerated design, control, and optimization.
- Inverse Design for Fluid-Structure Interactions using Graph Network Simulators. Advances in Neural Information Processing Systems 35 (2022), 13759–13774.
- José E Andrade and Utkarsh Mital. 2019. Multiscale and Multiphysics Modeling of Soils. In Geotechnical Fundamentals for Addressing New World Challenges. Springer, 141–168.
- Relational inductive biases, deep learning, and graph networks. arXiv preprint arXiv:1806.01261 (2018).
- Yongjin Choi and Krishna Kumar. 2023. Graph Neural Network-based surrogate model for granular flows. arXiv preprint arXiv:2305.05218 (2023).
- DoE. 2019. Workshop Report on Basic Research Needs for Scientific Machine Learning: Core Technologies for Artificial Intelligence. https://www.osti.gov/biblio/1478744
- A Haeri and K Skonieczny. 2021. Accurate and real-time simulation of rover wheel traction. In 2021 IEEE Aerospace Conference (50100). IEEE, 1–9.
- Difftaichi: Differentiable programming for physical simulation. arXiv preprint arXiv:1910.00935 (2019).
- Minority Report: A Graph Network Oracle for In Situ Visualization. arXiv preprint arXiv:2206.12683 (2022).
- Minority Report: A graph-network oracle for large-scale in situ visualization. In IEEE Large-Scale Data Analysis and Visualization. IEEE, Oklahoma, USA.
- Scalable and modular material point method for large-scale simulations. arXiv preprint arXiv:1909.13380 (2019).
- Krishna Kumar and Kenichi Soga. 2019. Large Deformation Modelling in Geomechanics. Geotechnical Design and Practice: Selected Topics (2019), 237–248.
- Krishna Kumar and Joseph Vantassel. 2022. GNS: A generalizable Graph Neural Network-based simulator for particulate and fluid modeling. arXiv preprint arXiv:2211.10228 (2022).
- Boundary graph neural networks for 3d simulations. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. 9099–9107.
- Utkarsh Mital and José E Andrade. 2021. Bridging length scales in granular materials using convolutional neural networks. Computational Particle Mechanics (2021), 1–15.
- Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational physics 378 (2019), 686–707.
- Learning to simulate complex physics with graph networks. In International Conference on Machine Learning. PMLR, 8459–8468.
- Samuel Schoenholz and Ekin Dogus Cubuk. 2020. Jax md: a framework for differentiable physics. Advances in Neural Information Processing Systems 33 (2020), 11428–11441.
- Trends in large-deformation analysis of landslide mass movements with particular emphasis on the material point method. Géotechnique 66, 3 (2016), 248–273.
- A cooperative game for automated learning of elasto-plasticity knowledge graphs and models with AI-guided experimentation. Computational Mechanics 64, 2 (2019), 467–499.
- Deep graph library: Towards efficient and scalable deep learning on graphs. arXiv preprint arXiv:1909.01315 (2019).
- Krishna Kumar (77 papers)
- Yongjin Choi (8 papers)