Physics-Informed Coarsening for Multigrid Graph Neural Surrogates
Amir Bazzi, David Cardinaux, Ramy Nemer, José Alves, Arjun Kalkur Matpadi Raghavendra, Elie Hachem · 2026
Read Physics-Informed Coarsening for Multigrid Graph Neural Surrogates by Amir Bazzi, David Cardinaux, Ramy Nemer, José Alves, Arjun Kalkur Matpadi Raghavendra, Elie Hachem (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-31013.txt.
Description
Bazzi and colleagues build a graph neural network surrogate for solids. Instead of sampling by shape rules, their coarsening scores nodes by local action. It keeps high strain or stress regions, putting power where it is most needed. This preserves distant links through layered message passing while raising stability over long rollouts. Tests across linear and transient regimes show steady gains in accuracy and control over sampling rivals. The results highlight this method for solid modeling.
Catalog record
| CREATOR | Amir Bazzi, David Cardinaux, Ramy Nemer, José Alves, Arjun Kalkur Matpadi Raghavendra, Elie Hachem |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49157 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.31013