Graph Fourier Neural ODEs: Bridging Spatial and Temporal Multiscales in Molecular Dynamics

Fang Sun,Zijie Huang,Haixin Wang,Yadi Cao,Xiao Luo,Wei Wang,Yizhou Sun

2024-11-03

Abstract:Molecular dynamics simulations are crucial for understanding complex physical, chemical, and biological processes at the atomic level. However, accurately capturing interactions across multiple spatial and temporal scales remains a significant challenge. We present a novel framework that jointly models spatial and temporal multiscale interactions in molecular dynamics. Our approach leverages Graph Fourier Transforms to decompose molecular structures into different spatial scales and employs Neural Ordinary Differential Equations to model the temporal dynamics in a curated manner influenced by the spatial modes. This unified framework links spatial structures with temporal evolution in a flexible manner, enabling more accurate and comprehensive simulations of molecular systems. We evaluate our model on the MD17 dataset, demonstrating consistent performance improvements over state-of-the-art baselines across multiple molecules, particularly under challenging conditions such as irregular timestep sampling and long-term prediction horizons. Ablation studies confirm the significant contributions of both spatial and temporal multiscale modeling components. Our method advances the simulation of complex molecular systems, potentially accelerating research in computational chemistry, drug discovery, and materials science.

Machine Learning,Chemical Physics,Quantitative Methods

What problem does this paper attempt to address?

The paper attempts to address the problem of accurately capturing interactions across multiple spatial and temporal scales in molecular dynamics simulations. Specifically, existing methods often focus on either the spatial or temporal scale, neglecting the intrinsic connections between them. This leads to limitations in the accuracy and comprehensiveness of simulations when dealing with complex molecular systems, especially under challenging conditions such as irregular time step sampling and long-term predictions.

To tackle this challenge, the authors propose a new framework that improves molecular dynamics simulations by jointly modeling multi-scale interactions in both space and time. This framework utilizes Graph Fourier Transform (GFT) to decompose molecular structures into different spatial scales and employs Neural Ordinary Differential Equations (Neural ODEs) to model the temporal dynamics influenced by these spatial patterns. This unified framework can flexibly link spatial structures with temporal evolution, enabling more accurate and comprehensive simulations of molecular systems.

The main contributions of the paper include:
1. **New multi-scale modeling perspective**: By explicitly linking spatial and temporal scales, it addresses the challenges of multi-scale modeling.
2. **Novel joint multi-scale modeling framework**: Combining GFT and Neural ODEs to effectively capture and propagate temporal dynamics across different spatial scales.
3. **Effective model architecture**: A new model architecture is designed and its effectiveness is validated through extensive experiments and ablation studies.

Through this approach, the authors hope to accelerate research progress in fields such as computational chemistry, drug discovery, and materials science.