MG-GNN: Multigrid Graph Neural Networks for Learning Multilevel Domain Decomposition Methods
Jul 24, 2023
Speakers
Ali Taghibakhshi
Speaker · 0 followers
Nicolas Nytko
Speaker · 0 followers
Tareq Zaman
Speaker · 0 followers
Scott MacLachlan
Speaker · 0 followers
Luke Olson
Speaker · 0 followers
Matthew West
Speaker · 0 followers
About
Domain decomposition methods (DDMs) are popular solvers for discretized systems of partial differential equations (PDEs), with one-level and multilevel variants. These solvers rely on several algorithmic and mathematical parameters, prescribing overlap, subdomain boundary conditions, and other properties of the DDM. While some work has been done on optimizing these parameters, it has mostly focused on the one-level setting or special cases such as structured-grid discretizations with regular subdomain construction. In this paper, we propose multigrid graph neural networks (MG-GNN), a novel GNN architecture for learning optimized parameters in two-level DDMs. We train MG-GNN using a new unsupervised loss function, enabling effective training on small problems that yields robust performance on unstructured grids that are orders of magnitude larger than those in the training set. We show that MG-GNN outperforms popular hierarchical graph network architectures for this optimization and that our proposed loss function is critical to achieving this improved performance.Domain decomposition methods (DDMs) are popular solvers for discretized systems of partial differential equations (PDEs), with one-level and multilevel variants. These solvers rely on several algorithmic and mathematical parameters, prescribing overlap, subdomain boundary conditions, and other properties of the DDM. While some work has been done on optimizing these parameters, it has mostly focused on the one-level setting or special cases such as structured-grid discretizations with regular sub…
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Who Needs to Know? Minimal Knowledge for Optimal Coordination
Watch later
Progressive Purification for Instance-Dependent Partial Label Learning
Watch later
Ning Xu, …
DRew: Dynamically Rewired Message Passing with Delay
Watch later
Closing Remarks
Watch later
Building a dialogue agent for Diplomacy
Watch later
