Graph neural network PDE solvers require at least a CFL-like number of message passes for hyperbolic problems and domain-spanning passes for parabolic and elliptic problems.
Learning mesh-based simulation with graph networks
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
support 1representative citing papers
citing papers explorer
-
On the under-reaching phenomenon in message-passing neural PDE solvers: revisiting the CFL condition
Graph neural network PDE solvers require at least a CFL-like number of message passes for hyperbolic problems and domain-spanning passes for parabolic and elliptic problems.