Research map / Graph learning / GNN architectures
Learning to Simulate Complex Physics with Graph Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| increase the number of message-passing steps (ablation) | GNS · Goop | rollout error | Helped |
| use unshared rather than shared processor GN parameters (ablation) | GNS · Goop | rollout accuracy | Helped |
| increase the connectivity radius (ablation) | GNS · Goop | rollout error | Helped |
| add an intermediate amount of input noise during training (ablation) | GNS · Goop | rollout accuracy | Helped |
| increase training noise scale (ablation) | GNS · Goop | one-step accuracy | Hurt |
| use relative rather than absolute positional encoding (ablation) | GNS · Goop | rollout error | Helped |
| condition on two recent velocities rather than one (ablation) | GNS · Goop | performance | Helped |
| increase history beyond two recent velocities (ablation) | GNS · Goop | performance | No clear effect |
| replace encoder and decoder MLPs with linear layers (ablation) | GNS · Goop | performance | No clear effect |
| remove self-edges from message passing (ablation) | GNS · Goop | performance | No clear effect |
| add global features and an updating global latent state (ablation) | GNS · Goop | performance | No clear effect |
| disable edge latent updates (ablation) | GNS · Goop | performance | No clear effect |
Closest papers by meaning
- MaNGO — Adaptable Graph Network Simulators via Meta-Learning (NeurIPS 2025 · similarity 0.92)
- Point Cloud Sequence Encoding for Material-conditioned Graph Network Simulators (NeurIPS 2026 · similarity 0.89)
- EvoMesh: Adaptive Physical Simulation with Hierarchical Graph Evolutions (ICML 2025 · similarity 0.88)
- Diffusion-Based Hierarchical Graph Neural Networks for Simulating Nonlinear Solid Mechanics (NeurIPS 2025 · similarity 0.88)
- Physics-Informed DeepONets for drift-diffusion on metric graphs: simulation and parameter identification (ICML 2025 · similarity 0.87)
- Graph Mamba Operator: A Latent Simulator for Interacting Particle Systems (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.