Research map / Time series / Long-term forecasting

Non-stationary Equivariant Graph Neural Networks for Physical Dynamics Simulation

NeurIPS 2025 · Time series › Long-term forecasting · approach: Spatiotemporal forecasting models · 2 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a non-stationary equivariant GNN for physical dynamics simulation that accounts for the time-varying joint distribution of dynamics.

Problem

Existing equivariant GNNs for physical dynamics ignore the non-stationarity of dynamics, where the joint distribution changes over time, limiting generalization.

Approach

Extends equivariant GNNs with mechanisms for modeling non-stationary physical dynamics, rather than assuming a stationary joint distribution; the abstract's critique of prior non-stationary approaches suggests the method addresses their shortcomings.

Closest papers by meaning

• Equivariant Spatio-Temporal Attentive Graph Networks to Simulate Physical Dynamics (NeurIPS 2023 · similarity 0.92)
• SEGNO: Generalizing Equivariant Graph Neural Networks with Physical Inductive Biases (ICLR 2024 · similarity 0.90)
• Equivariant Graph Neural Operator for Modeling 3D Dynamics (ICML 2024 · similarity 0.89)
• Universally Invariant Learning in Equivariant GNNs (NeurIPS 2025 · similarity 0.85)
• Long-Short-Range Message-Passing: A Physics-Informed Framework to Capture Non-Local Interaction for Scalable Molecular Dynamics Simulation (ICLR 2024 · similarity 0.85)
• Are High-Degree Representations Really Unnecessary in Equivariant Graph Neural Networks? (NeurIPS 2024 · similarity 0.85)