Prometheus: Out-of-distribution Fluid Dynamics Modeling with Disentangled Graph ODE

Hao Wu, Huiyuan Wang, Kun Wang, Weiyan Wang, Changan Ye, Yangyu Tao, Chong Chen, Xian-Sheng Hua, Xiao Luo
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:53870-53891, 2024.

Abstract

Fluid dynamics modeling has received extensive attention in the machine learning community. Although numerous graph neural network (GNN) approaches have been proposed for this problem, the problem of out-of-distribution (OOD) generalization remains underexplored. In this work, we propose a new large-scale dataset Prometheus which simulates tunnel and pool fires across various environmental conditions and builds an extensive benchmark of 12 baselines, which demonstrates that the OOD generalization performance is far from satisfactory. To tackle this, this paper introduces a new approach named Disentangled Graph ODE (DGODE), which learns disentangled representations for continuous interacting dynamics modeling. In particular, we utilize a temporal GNN and a frequency network to extract semantics from historical trajectories into node representations and environment representations respectively. To mitigate the potential distribution shift, we minimize the mutual information between invariant node representations and the discretized environment features using adversarial learning. Then, they are fed into a coupled graph ODE framework, which models the evolution using neighboring nodes and dynamical environmental context. In addition, we enhance the stability of the framework by perturbing the environment features to enhance robustness. Extensive experiments validate the effectiveness of DGODE compared with state-of-the-art approaches.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-wu24aa, title = {Prometheus: Out-of-distribution Fluid Dynamics Modeling with Disentangled Graph {ODE}}, author = {Wu, Hao and Wang, Huiyuan and Wang, Kun and Wang, Weiyan and Ye, Changan and Tao, Yangyu and Chen, Chong and Hua, Xian-Sheng and Luo, Xiao}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {53870--53891}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/wu24aa/wu24aa.pdf}, url = {https://proceedings.mlr.press/v235/wu24aa.html}, abstract = {Fluid dynamics modeling has received extensive attention in the machine learning community. Although numerous graph neural network (GNN) approaches have been proposed for this problem, the problem of out-of-distribution (OOD) generalization remains underexplored. In this work, we propose a new large-scale dataset Prometheus which simulates tunnel and pool fires across various environmental conditions and builds an extensive benchmark of 12 baselines, which demonstrates that the OOD generalization performance is far from satisfactory. To tackle this, this paper introduces a new approach named Disentangled Graph ODE (DGODE), which learns disentangled representations for continuous interacting dynamics modeling. In particular, we utilize a temporal GNN and a frequency network to extract semantics from historical trajectories into node representations and environment representations respectively. To mitigate the potential distribution shift, we minimize the mutual information between invariant node representations and the discretized environment features using adversarial learning. Then, they are fed into a coupled graph ODE framework, which models the evolution using neighboring nodes and dynamical environmental context. In addition, we enhance the stability of the framework by perturbing the environment features to enhance robustness. Extensive experiments validate the effectiveness of DGODE compared with state-of-the-art approaches.} }
Endnote
%0 Conference Paper %T Prometheus: Out-of-distribution Fluid Dynamics Modeling with Disentangled Graph ODE %A Hao Wu %A Huiyuan Wang %A Kun Wang %A Weiyan Wang %A Changan Ye %A Yangyu Tao %A Chong Chen %A Xian-Sheng Hua %A Xiao Luo %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-wu24aa %I PMLR %P 53870--53891 %U https://proceedings.mlr.press/v235/wu24aa.html %V 235 %X Fluid dynamics modeling has received extensive attention in the machine learning community. Although numerous graph neural network (GNN) approaches have been proposed for this problem, the problem of out-of-distribution (OOD) generalization remains underexplored. In this work, we propose a new large-scale dataset Prometheus which simulates tunnel and pool fires across various environmental conditions and builds an extensive benchmark of 12 baselines, which demonstrates that the OOD generalization performance is far from satisfactory. To tackle this, this paper introduces a new approach named Disentangled Graph ODE (DGODE), which learns disentangled representations for continuous interacting dynamics modeling. In particular, we utilize a temporal GNN and a frequency network to extract semantics from historical trajectories into node representations and environment representations respectively. To mitigate the potential distribution shift, we minimize the mutual information between invariant node representations and the discretized environment features using adversarial learning. Then, they are fed into a coupled graph ODE framework, which models the evolution using neighboring nodes and dynamical environmental context. In addition, we enhance the stability of the framework by perturbing the environment features to enhance robustness. Extensive experiments validate the effectiveness of DGODE compared with state-of-the-art approaches.
APA
Wu, H., Wang, H., Wang, K., Wang, W., Ye, C., Tao, Y., Chen, C., Hua, X. & Luo, X.. (2024). Prometheus: Out-of-distribution Fluid Dynamics Modeling with Disentangled Graph ODE. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:53870-53891 Available from https://proceedings.mlr.press/v235/wu24aa.html.

Related Material