mlsanityData to 14 Oct 2026

Force-Free Molecular Dynamics Through Autoregressive Equivariant Networks

arXiv 2503.23794 ↗First posted 31 Mar 2025physics.comp-ph cond-mat.mtrl-sci cs.LGcited by 15

Not on the map. Why?

Abstract (arXiv)

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× 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.

Key facts. Cited by 15 papers in the map (data through 14 Oct 2026). Builds on 28 of its 111 references that are in the map. First posted on arXiv 31 Mar 2025.

Builds on

28 of 111 references.

References · 28

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  2. 2017Neural Message Passing for Quantum Chemistrycited by 3,487
  3. 2019On the Convergence of Adam and Beyondcited by 841
  4. 2018Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point cloudscited by 746
  5. 2021E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentialscited by 708
  6. 2017Deep Learning Scaling is Predictable, Empiricallycited by 582
  7. 2017SchNet: A continuous-filter convolutional neural network for modeling quantum interactionscited by 571
  8. 2020Directional Message Passing for Molecular Graphscited by 561
  9. 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 474
  10. 20183D Steerable CNNs: Learning Rotationally Equivariant Features in Volumetric Datacited by 354
  11. 2022Learning Local Equivariant Representations for Large-Scale Atomistic Dynamicscited by 256
  12. 2021Geometric and Physical Quantities Improve E(3) Equivariant Message Passingcited by 228
16 more
  1. 2020Machine Learning Force Fieldscited by 207
  2. 2022e3nn: Euclidean Neural Networkscited by 174
  3. 2021SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effectscited by 104
  4. 2022The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentialscited by 101
  5. 2021Learned Coarse Models for Efficient Turbulence Simulationcited by 91
  6. 2023Timewarp: Transferable Acceleration of Molecular Dynamics by Learning Time-Coarsened Dynamicscited by 66
  7. 2024Generative Modeling of Molecular Dynamics Trajectoriescited by 51
  8. 2024Equivariant Graph Neural Operator for Modeling 3D Dynamicscited by 38
  9. 2023Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic sizecited by 33
  10. 2023Evaluation of the MACE Force Field Architecture: from Medicinal Chemistry to Materials Sciencecited by 25
  11. 2022Simulate Time-integrated Coarse-grained Molecular Dynamics with Multi-Scale Graph Networkscited by 23
  12. 2024Flow Matching for Accelerated Simulation of Atomic Transport in Crystalline Materialscited by 9
  13. 2020Solving Newton's Equations of Motion with Large Timesteps using Recurrent Neural Networks based Operatorscited by 6
  14. 2023Learning Interatomic Potentials at Multiple Scalescited by 5
  15. 2022Super-resolution in Molecular Dynamics Trajectory Reconstruction with Bi-Directional Neural Networkscited by 3
  16. 2024BoostMD: Accelerating molecular sampling by leveraging ML force field features from previous time-stepscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 6; 2026: 9.

15 papers.

Sort:

By area:

  1. 2025FlashMD: long-stride, universal prediction of molecular dynamicscited by 12
  2. 2025Transferable Generative Models Bridge Femtosecond to Nanosecond Time-Step Molecular Dynamicscited by 11
  3. 2026Learning Hamiltonian Flow Maps: Mean Flow Consistency for Large-Timestep Molecular Dynamicscited by 7
  4. 2025HollowFlow: Efficient Sample Likelihood Evaluation using Hollow Message Passingcited by 5
  5. 2026JEPA-Anything: Learning Predictive Models across Different Worldscited by 3
  6. 2025Learning the action for long-time-step simulations of molecular dynamicscited by 2
  7. 2025PAINET: A Principled Efficient Transformer for 3D Dynamics Modelingcited by 2
  8. 2025ATOM: A Pretrained Neural Operator for Multitask Molecular Dynamicscited by 1
  9. 2026Generative Pseudo-Force Fields for Molecular Generationcited by 1
  10. 2026Implicit Machine Learning Force Fields Accelerate Molecular Dynamics Simulationscited by 1
  11. 2026A Priori Sampling of Transition States with Guided Diffusioncited by 0
  12. 2026Orthogonal JEPA: Factorized Predictive States for Latent World Modelscited by 0
3 more
  1. 2026SupraTITO: Transferable Generative Molecular Dynamics for Supramolecular Systemscited by 0
  2. 2026Langevin Flow Maps: Efficient Molecular Dynamics and Transition Path Samplingcited by 0
  3. 2026Looking Under the Streetlight: Evaluation in Generative Molecular Dynamicscited by 0