Redirecting to original paper in 30 seconds...
Click below to go immediately or wait for automatic redirect
📄 Abstract
Abstract: Molecular dynamics (MD) is a powerful tool for exploring the behavior of
atomistic systems, but its reliance on sequential numerical integration limits
simulation efficiency. We present a novel neural network architecture,
MDtrajNet, and a pre-trained foundational model, MDtrajNet-1, that directly
generates MD trajectories across chemical space, bypassing force calculations
and integration. This approach accelerates simulations by up to two orders of
magnitude compared to traditional MD, even those enhanced by machine-learning
interatomic potentials. MDtrajNet combines equivariant neural networks with a
transformer-based architecture to achieve strong accuracy and transferability
in predicting long-time trajectories. Remarkably, the errors of the
trajectories generated by MDtrajNet-1 for various known and unseen molecular
systems are close to those of the conventional ab initio MD. The architecture's
flexible design supports diverse application scenarios, including different
statistical ensembles, boundary conditions, and interaction types. By
overcoming the intrinsic speed barrier of conventional MD, MDtrajNet opens new
frontiers in efficient and scalable atomistic simulations.
arXiv Category
physics.chem-ph
Key Contributions
Introduces MDtrajNet, a novel neural network architecture and foundational model that directly generates molecular dynamics trajectories, bypassing traditional force calculations and integration. This approach achieves up to two orders of magnitude acceleration compared to conventional MD methods, with accuracy comparable to ab initio MD, enabling efficient exploration of chemical space.
Business Value
Dramatically accelerates drug discovery, materials design, and understanding of chemical processes by enabling faster and more extensive molecular simulations.