Force-Free Molecular Dynamics Through Autoregressive Equivariant Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Thiemann, Fabian L., Reschützegger, Thiago, Esposito, Massimiliano, Taddese, Tseden, Olarte-Plata, Juan D., Martelli, Fausto
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917972428193792
author Thiemann, Fabian L.
Reschützegger, Thiago
Esposito, Massimiliano
Taddese, Tseden
Olarte-Plata, Juan D.
Martelli, Fausto
author_facet Thiemann, Fabian L.
Reschützegger, Thiago
Esposito, Massimiliano
Taddese, Tseden
Olarte-Plata, Juan D.
Martelli, Fausto
contents Molecular dynamics (MD) simulations play a crucial role in scientific research. Yet their computational cost often limits the timescales and system sizes that can be explored. Most data-driven efforts have been focused on reducing the computational cost of accurate interatomic forces required for solving the equations of motion. Despite their success, however, these machine learning interatomic potentials (MLIPs) are still bound to small time-steps. In this work, we introduce TrajCast, a transferable and data-efficient framework based on autoregressive equivariant message passing networks that directly updates atomic positions and velocities lifting the constraints imposed by traditional numerical integration. We benchmark our framework across various systems, including a small molecule, crystalline material, and bulk liquid, demonstrating excellent agreement with reference MD simulations for structural, dynamical, and energetic properties. Depending on the system, TrajCast allows for forecast intervals up to $30\times$ larger than traditional MD time-steps, generating over 15 ns of trajectory data per day for a solid with more than 4,000 atoms. By enabling efficient large-scale simulations over extended timescales, TrajCast can accelerate materials discovery and explore physical phenomena beyond the reach of traditional simulations and experiments. An open-source implementation of TrajCast is accessible under https://github.com/IBM/trajcast.
format Preprint
id arxiv_https___arxiv_org_abs_2503_23794
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Force-Free Molecular Dynamics Through Autoregressive Equivariant Networks
Thiemann, Fabian L.
Reschützegger, Thiago
Esposito, Massimiliano
Taddese, Tseden
Olarte-Plata, Juan D.
Martelli, Fausto
Computational Physics
Materials Science
Machine Learning
Chemical Physics
Molecular dynamics (MD) simulations play a crucial role in scientific research. Yet their computational cost often limits the timescales and system sizes that can be explored. Most data-driven efforts have been focused on reducing the computational cost of accurate interatomic forces required for solving the equations of motion. Despite their success, however, these machine learning interatomic potentials (MLIPs) are still bound to small time-steps. In this work, we introduce TrajCast, a transferable and data-efficient framework based on autoregressive equivariant message passing networks that directly updates atomic positions and velocities lifting the constraints imposed by traditional numerical integration. We benchmark our framework across various systems, including a small molecule, crystalline material, and bulk liquid, demonstrating excellent agreement with reference MD simulations for structural, dynamical, and energetic properties. Depending on the system, TrajCast allows for forecast intervals up to $30\times$ larger than traditional MD time-steps, generating over 15 ns of trajectory data per day for a solid with more than 4,000 atoms. By enabling efficient large-scale simulations over extended timescales, TrajCast can accelerate materials discovery and explore physical phenomena beyond the reach of traditional simulations and experiments. An open-source implementation of TrajCast is accessible under https://github.com/IBM/trajcast.
title Force-Free Molecular Dynamics Through Autoregressive Equivariant Networks
topic Computational Physics
Materials Science
Machine Learning
Chemical Physics
url https://arxiv.org/abs/2503.23794