# Learning to Simulate Complex Physics with Graph Networks

*Source* · `src-sanchezgonzalez-2020-gns` · status: **n/a**

| field | value |
|---|---|
| kind | Peer-reviewed paper |
| authors | Alvaro Sanchez-Gonzalez, Jonathan Godwin, Tobias Pfaff, Rex Ying, Jure Leskovec, Peter W. Battaglia |
| venue | ICML 2020 |
| date | 2020-02-21 |
| url | https://arxiv.org/abs/2002.09405 |
| arxiv | 2002.09405 |
| accessed | 2026-10-24 |

## Notes

GNS: particles as graph nodes with learned message passing for fluids, rigid solids, deformables; trained on one-step targets, generalizes to thousands of steps and 10x more particles; training noise limits drift.

## Backlinks (computed)

**cited by:** [Graph network simulators and neural operators](../technique/graph-network-simulators.md)

## Citation

Alvaro Sanchez-Gonzalez, Jonathan Godwin, Tobias Pfaff, Rex Ying, Jure Leskovec, Peter W. Battaglia (2020-02-21). *Learning to Simulate Complex Physics with Graph Networks*. ICML 2020 https://arxiv.org/abs/2002.09405

---
Provenance: created in run `2026-10-24-techniques-sim-math`.