Multi-Scale Physical Representations for Approximating PDE Solutions with Graph Neural Operators.
Léon MigusYuan YinJocelyn Ahmed MazariPatrick GallinariPublished in: TAG-ML (2022)
Keyphrases
- multiscale
- partial differential equations
- graph structure
- neural network
- distributed representations
- edge detection
- graph theory
- graph representation
- directed graph
- solution space
- level set
- graph model
- numerical algorithms
- natural images
- anisotropic diffusion
- weighted graph
- graph representations
- morphological operators
- differential operators
- image magnification
- connected components
- image analysis
- laplacian operator
- neural model
- simulated annealing
- graph mining
- associative memory
- random walk
- structured data