GNN-Surrogate: A Hierarchical and Adaptive Graph Neural Network for Parameter Space Exploration of Unstructured-Mesh Ocean Simulations.
Neng ShiJiayi XuSkylar W. WursterHanqi GuoJonathan WoodringLuke P. Van RoekelHan-Wei ShenPublished in: CoRR (2022)
Keyphrases
- space exploration
- neural network
- structured data
- artificial neural networks
- graph structure
- random walk
- obstacle avoidance
- neural fuzzy
- back propagation
- graph representation
- graph model
- multi layer
- genetic algorithm
- neural network model
- bp neural network
- topological information
- numerical simulations
- directed graph
- graph theory
- parameter values
- graph mining
- real time
- semi structured
- hierarchical structure
- multi modal
- neural network is trained
- mesh refinement