Papers
Topics
Authors
Recent
Search
2000 character limit reached

MAgNET: A Graph U-Net Architecture for Mesh-Based Simulations

Published 1 Nov 2022 in cs.LG and cs.CE | (2211.00713v3)

Abstract: In many cutting-edge applications, high-fidelity computational models prove to be too slow for practical use and are therefore replaced by much faster surrogate models. Recently, deep learning techniques have increasingly been utilized to accelerate such predictions. To enable learning on large-dimensional and complex data, specific neural network architectures have been developed, including convolutional and graph neural networks. In this work, we present a novel encoder-decoder geometric deep learning framework called MAgNET, which extends the well-known convolutional neural networks to accommodate arbitrary graph-structured data. MAgNET consists of innovative Multichannel Aggregation (MAg) layers and graph pooling/unpooling layers, forming a graph U-Net architecture that is analogous to convolutional U-Nets. We demonstrate the predictive capabilities of MAgNET in surrogate modeling for non-linear finite element simulations in the mechanics of solids.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (75)
  1. A mechanics-informed artificial neural network approach in data-driven constitutive modeling. doi:10.2514/6.2022-0100.
  2. General multi-fidelity framework for training artificial neural networks with computational models. Frontiers in Materials 6, 61.
  3. Spectral clustering with graph neural networks for graph pooling. URL: https://arxiv.org/abs/1907.00481, doi:10.48550/ARXIV.1907.00481.
  4. Learning finite element convergence with the multi-fidelity graph neural network. Computer Methods in Applied Mechanics and Engineering 397, 115120. URL: https://www.sciencedirect.com/science/article/pii/S004578252200305X.
  5. Finite Element Multigrid Methods. Springer New York, New York, NY. pp. 155–173. doi:10.1007/978-0-387-75934-0_7.
  6. Geometric deep learning: Going beyond euclidean data. IEEE Signal Processing Magazine 34, 18–42. doi:10.1109/MSP.2017.2693418.
  7. Physics-based deep neural network for augmented reality during liver surgery, in: Shen, D., Liu, T., Peters, T.M., Staib, L.H., Essert, C., Zhou, S., Yap, P.T., Khan, A. (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2019, Springer International Publishing, Cham. pp. 137–145.
  8. Real-time error control for surgical simulation. IEEE Transactions on Biomedical Engineering 65, 596–607. doi:10.1109/TBME.2017.2695587.
  9. Graph coarsening with neural networks, in: International Conference on Learning Representations. URL: https://openreview.net/forum?id=uxpzitPEooJ.
  10. Simple and deep graph convolutional networks, in: III, H.D., Singh, A. (Eds.), Proceedings of the 37th International Conference on Machine Learning, PMLR. pp. 1725–1735. URL: https://proceedings.mlr.press/v119/chen20v.html.
  11. 3d u-net: learning dense volumetric segmentation from sparse annotation, in: International conference on medical image computing and computer-assisted intervention, Springer. pp. 424–432.
  12. Model order reduction assisted by deep neural networks (rom-net). Advanced Modeling and Simulation in Engineering Sciences 7, 1–27.
  13. Magnet: A graph U-Net architecture for mesh-based simulations [dataset]. doi:10.5281/zenodo.000000.
  14. Probabilistic deep learning for real-time large deformation simulations. Computer Methods in Applied Mechanics and Engineering 398, 115307. URL: https://www.sciencedirect.com/science/article/pii/S004578252200411X, doi:https://doi.org/10.1016/j.cma.2022.115307.
  15. Convolution, aggregation and attention based deep neural networks for accelerating simulations in mechanics. Frontiers in Materials 10. URL: https://doi.org/10.3389/fmats.2023.1128954, doi:10.3389/fmats.2023.1128954.
  16. On the development of volumetric strain energy functions. J. Appl. Mech. 67, 17–21.
  17. A deep encoder-decoder for surrogate modelling of liquid moulding of composites. Engineering Applications of Artificial Intelligence 120, 105945. URL: https://www.sciencedirect.com/science/article/pii/S095219762300129X, doi:https://doi.org/10.1016/j.engappai.2023.105945.
  18. Graph u-net. URL: https://openreview.net/forum?id=HJePRoAct7.
  19. Phygeonet: Physics-informed geometry-adaptive convolutional neural networks for solving parameterized steady-state pdes on irregular domain. Journal of Computational Physics 428, 110079. URL: https://www.sciencedirect.com/science/article/pii/S0021999120308536.
  20. Physics-informed graph neural galerkin networks: A unified framework for solving pde-governed forward and inverse problems. Computer Methods in Applied Mechanics and Engineering 390, 114502. URL: https://www.sciencedirect.com/science/article/pii/S0045782521007076, doi:https://doi.org/10.1016/j.cma.2021.114502.
  21. A multi-fidelity wind surface pressure assessment via machine learning: A high-rise building case. Building and Environment 234, 110135. URL: https://www.sciencedirect.com/science/article/pii/S0360132323001622, doi:https://doi.org/10.1016/j.buildenv.2023.110135.
  22. Fast, generic, and reliable control and simulation of soft robots using model order reduction. IEEE Transactions on Robotics 34, 1565–1576.
  23. Physics informed neural networks for continuum micromechanics. Computer Methods in Applied Mechanics and Engineering 393, 114790. URL: https://www.sciencedirect.com/science/article/pii/S0045782522001268, doi:https://doi.org/10.1016/j.cma.2022.114790.
  24. Spleeter: a fast and efficient music source separation tool with pre-trained models. Journal of Open Source Software 5, 2154. doi:10.21105/joss.02154.
  25. Data-driven methods for stress field predictions in random heterogeneous materials. Engineering Applications of Artificial Intelligence 123, 106267. URL: https://www.sciencedirect.com/science/article/pii/S0952197623004517, doi:https://doi.org/10.1016/j.engappai.2023.106267.
  26. Learning constitutive relations from indirect observations using deep neural networks. Journal of Computational Physics 416, 109491. URL: https://www.sciencedirect.com/science/article/pii/S0021999120302655, doi:https://doi.org/10.1016/j.jcp.2020.109491.
  27. Graph neural networks (gnns) based accelerated numerical simulation. Engineering Applications of Artificial Intelligence 123, 106370. URL: https://www.sciencedirect.com/science/article/pii/S0952197623005547, doi:https://doi.org/10.1016/j.engappai.2023.106370.
  28. Niftysim: A gpu-based nonlinear finite element package for simulation of soft tissue biomechanics. International journal of computer assisted radiology and surgery 10, 1077–1095.
  29. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 .
  30. Machine learning-accelerated computational fluid dynamics. Proceedings of the National Academy of Sciences 118, e2101784118. URL: https://www.pnas.org/doi/abs/10.1073/pnas.2101784118, doi:10.1073/pnas.2101784118, arXiv:https://www.pnas.org/doi/pdf/10.1073/pnas.2101784118.
  31. Multi-language and multi-environment generation of nonlinear finite element codes. Engineering with Computers 18, 312--327. doi:10.1007/s003660200028.
  32. A graph-based probabilistic geometric deep learning framework with online enforcement of physical constraints to predict the criticality of defects in porous materials. International Journal of Solids and Structures 286-287, 112545. URL: https://www.sciencedirect.com/science/article/pii/S0020768323004420, doi:https://doi.org/10.1016/j.ijsolstr.2023.112545.
  33. A bayesian multiscale cnn framework to predict local stress fields in structures with microscale features. Computational Mechanics 69, 733--766. doi:10.1007/s00466-021-02112-3.
  34. Identification of material parameters and traction field for soft bodies in contact. Computer Methods in Applied Mechanics and Engineering 406, 115889. URL: https://www.sciencedirect.com/science/article/pii/S0045782523000129, doi:https://doi.org/10.1016/j.cma.2023.115889.
  35. Digital volume correlation for large deformations of soft tissues: Pipeline and proof of concept for the application to breast ex vivo deformations. Journal of the Mechanical Behavior of Biomedical Materials , 105490URL: https://www.sciencedirect.com/science/article/pii/S1751616122003952, doi:https://doi.org/10.1016/j.jmbbm.2022.105490.
  36. Towards incompressible laminar flow estimation based on interpolated feature generation and deep learning. Sustainability 14. URL: https://www.mdpi.com/2071-1050/14/19/11996, doi:10.3390/su141911996.
  37. Self-attention graph pooling. URL: https://arxiv.org/abs/1904.08082, doi:10.48550/ARXIV.1904.08082.
  38. Neural extended kalman filters for learning and predicting dynamics of structural systems. Structural Health Monitoring , 14759217231179912URL: https://doi.org/10.1177/14759217231179912, doi:10.1177/14759217231179912, arXiv:2210.04165.
  39. A framework for modelling the biomechanical behaviour of the human liver during breathing in real time using machine learning. Expert Systems with Applications 71, 342--357. URL: https://www.sciencedirect.com/science/article/pii/S0957417416306728, doi:https://doi.org/10.1016/j.eswa.2016.11.037.
  40. Clique pooling for graph classification. URL: https://arxiv.org/abs/1904.00374, doi:10.48550/ARXIV.1904.00374.
  41. Deep learning acceleration of total lagrangian explicit dynamics for soft tissue mechanics. Computer Methods in Applied Mechanics and Engineering 358, 112628. URL: https://www.sciencedirect.com/science/article/pii/S0045782519305109, doi:https://doi.org/10.1016/j.cma.2019.112628.
  42. Simulation of hyperelastic materials in real-time using deep learning. Medical Image Analysis 59, 101569. doi:10.1016/j.media.2019.101569.
  43. Deep learning predicts path-dependent plasticity. Proceedings of the National Academy of Sciences 116, 26414--26420. URL: https://doi.org/10.1073/pnas.1911815116, doi:10.1073/pnas.1911815116.
  44. Physics-informed neural networks for non-newtonian fluid thermo-mechanical problems: An application to rubber calendering process. Engineering Applications of Artificial Intelligence 114, 105176. URL: https://www.sciencedirect.com/science/article/pii/S0952197622002810, doi:https://doi.org/10.1016/j.engappai.2022.105176.
  45. Non-intrusive surrogate modeling for parametrized time-dependent partial differential equations using convolutional autoencoders. Engineering Applications of Artificial Intelligence 109, 104652. URL: https://www.sciencedirect.com/science/article/pii/S0952197621004541, doi:https://doi.org/10.1016/j.engappai.2021.104652.
  46. Cfdnet: A deep learning-based accelerator for fluid simulations, Association for Computing Machinery, New York, NY, USA. URL: https://doi.org/10.1145/3392717.3392772, doi:10.1145/3392717.3392772.
  47. Deepphysics: a physics aware deep learning framework for real-time simulation. URL: https://arxiv.org/abs/2109.09491, doi:10.48550/ARXIV.2109.09491.
  48. Non-linear elastic deformations. Dover Publications.
  49. Deep learning for reduced order modelling and efficient temporal evolution of fluid simulations. Physics of Fluids 33, 107101. URL: https://doi.org/10.1063%2F5.0062546, doi:10.1063/5.0062546.
  50. Real-time biomechanical modeling of the liver using machine learning models trained on finite element method simulations. Expert Systems with Applications 143, 113083. URL: https://www.sciencedirect.com/science/article/pii/S0957417419308000, doi:https://doi.org/10.1016/j.eswa.2019.113083.
  51. Learning mesh-based simulation with graph networks, in: International Conference on Learning Representations. URL: https://openreview.net/forum?id=roNqYL0_XP.
  52. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics 378, 686--707. URL: https://www.sciencedirect.com/science/article/pii/S0021999118307125, doi:https://doi.org/10.1016/j.jcp.2018.10.045.
  53. Three-dimensional convolutional neural network (3d-cnn) for heterogeneous material homogenization. Computational Materials Science 184, 109850. doi:10.1016/j.commatsci.2020.109850.
  54. A causal u-net based neural beamforming network for real-time multi-channel speech enhancement, pp. 1832--1836. doi:10.21437/Interspeech.2021-1457.
  55. U-net: Convolutional networks for biomedical image segmentation, in: Medical Image Computing and Computer-Assisted Intervention (MICCAI), Springer. pp. 234--241. URL: http://lmb.informatik.uni-freiburg.de/Publications/2015/RFB15a. (available on arXiv:1505.04597 [cs.CV]).
  56. A data-driven physics-constrained deep learning computational framework for solving von mises plasticity. Engineering Applications of Artificial Intelligence 122, 106049. URL: https://www.sciencedirect.com/science/article/pii/S0952197623002336, doi:https://doi.org/10.1016/j.engappai.2023.106049.
  57. Fem-based training of artificial neural networks for modular soft robots, in: 2017 IEEE International Conference on Robotics and Biomimetics (ROBIO), pp. 385--392. doi:10.1109/ROBIO.2017.8324448.
  58. Design, fabrication and control of soft robots. Nature 521, 467--475.
  59. An energy approach to the solution of partial differential equations in computational mechanics via machine learning: Concepts, implementation and applications. Computer Methods in Applied Mechanics and Engineering 362, 112790. URL: https://www.sciencedirect.com/science/article/pii/S0045782519306826, doi:https://doi.org/10.1016/j.cma.2019.112790.
  60. Learning to simulate complex physics with graph networks. Learning to simulate complex physics with graph networks , 8459 – 8468URL: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85094802982&partnerID=40&md5=57a515ee79e6915a2266fa3f2bc4870c. cited by: 53.
  61. Graph neural networks and implicit neural representation for near-optimal topology prediction over irregular design domains. Engineering Applications of Artificial Intelligence 123, 106284. URL: https://www.sciencedirect.com/science/article/pii/S0952197623004682, doi:https://doi.org/10.1016/j.engappai.2023.106284.
  62. Graph neural network-based surrogate models for finite element analysis arXiv:2211.09373.
  63. Penalty function formulations for incompressible nonlinear elastostatics. Computer Methods in Applied Mechanics and Engineering 35, 107--118.
  64. Nn-euclid: Deep-learning hyperelasticity without stress data. Journal of the Mechanics and Physics of Solids 169, 105076. URL: https://www.sciencedirect.com/science/article/pii/S0022509622002538, doi:https://doi.org/10.1016/j.jmps.2022.105076.
  65. Management of an academic hpc cluster: The ul experience. doi:10.1109/HPCSim.2014.6903792.
  66. Graph attention networks. URL: https://arxiv.org/abs/1710.10903, doi:10.48550/ARXIV.1710.10903.
  67. Neural-network acceleration of projection-based model-order-reduction for finite plasticity: Application to RVEs. arXiv URL: https://arxiv.org/abs/2109.07747, doi:10.48550/ARXIV.2109.07747.
  68. Geometric deep learning for computational mechanics part i: anisotropic hyperelasticity. Computer Methods in Applied Mechanics and Engineering 371, 113299. URL: https://www.sciencedirect.com/science/article/pii/S0045782520304849, doi:https://doi.org/10.1016/j.cma.2020.113299.
  69. Multi-resolution convolutional neural networks for inverse problems. Scientific Reports 10, 5730. doi:10.1038/s41598-020-62484-z.
  70. A comprehensive survey on graph neural networks. IEEE Transactions on Neural Networks and Learning Systems 32, 4--24. doi:10.1109/TNNLS.2020.2978386.
  71. Spatial temporal graph convolutional networks for skeleton-based action recognition. URL: https://arxiv.org/abs/1801.07455, doi:10.48550/ARXIV.1801.07455.
  72. Multi-fidelity surrogate modeling for temperature field prediction using deep convolution neural network. Engineering Applications of Artificial Intelligence 123, 106354. URL: https://www.sciencedirect.com/science/article/pii/S0952197623005389, doi:https://doi.org/10.1016/j.engappai.2023.106354.
  73. Semantic graph convolutional networks for 3d human pose regression, in: 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE. URL: https://doi.org/10.1109%2Fcvpr.2019.00354, doi:10.1109/cvpr.2019.00354.
  74. Physics-informed convolutional neural networks for temperature field prediction of heat source layout without labeled data. Engineering Applications of Artificial Intelligence 117, 105516. URL: https://www.sciencedirect.com/science/article/pii/S0952197622005061, doi:https://doi.org/10.1016/j.engappai.2022.105516.
  75. Graph neural networks: A review of methods and applications. AI Open 1, 57--81. URL: https://www.sciencedirect.com/science/article/pii/S2666651021000012, doi:https://doi.org/10.1016/j.aiopen.2021.01.001.
Citations (21)
View on Semantic Scholar

Summary

No one has generated a summary of this paper yet.

Sign Up to Summarize

Paper to Video (Beta)

No one has generated a video about this paper yet.

Sign Up to Generate All Videos Subscribe on YouTube

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Sign Up to Generate

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Generate Now

Continue Learning

We haven't generated follow-up questions for this paper yet.

Generate Now

Collections

Sign up for free to add this paper to one or more collections.

Sign Up