Learning Local Equivariant Representations for Large-Scale Atomistic Dynamics
This work addresses the problem of balancing accuracy and computational efficiency in atomistic dynamics for researchers in computational chemistry and materials science, offering a novel approach that is not incremental but introduces a new paradigm.
The paper tackles the challenge of creating accurate and scalable interatomic potentials for molecular and materials simulations by introducing Allegro, a strictly local equivariant deep learning method that avoids message passing. It achieves improvements over state-of-the-art methods on QM9 and MD-17 datasets, with a single layer outperforming deep message passing networks and transformers, and demonstrates scalability to 100 million atoms in simulations.
A simultaneously accurate and computationally efficient parametrization of the energy and atomic forces of molecules and materials is a long-standing goal in the natural sciences. In pursuit of this goal, neural message passing has lead to a paradigm shift by describing many-body correlations of atoms through iteratively passing messages along an atomistic graph. This propagation of information, however, makes parallel computation difficult and limits the length scales that can be studied. Strictly local descriptor-based methods, on the other hand, can scale to large systems but do not currently match the high accuracy observed with message passing approaches. This work introduces Allegro, a strictly local equivariant deep learning interatomic potential that simultaneously exhibits excellent accuracy and scalability of parallel computation. Allegro learns many-body functions of atomic coordinates using a series of tensor products of learned equivariant representations, but without relying on message passing. Allegro obtains improvements over state-of-the-art methods on the QM9 and revised MD-17 data sets. A single tensor product layer is shown to outperform existing deep message passing neural networks and transformers on the QM9 benchmark. Furthermore, Allegro displays remarkable generalization to out-of-distribution data. Molecular dynamics simulations based on Allegro recover structural and kinetic properties of an amorphous phosphate electrolyte in excellent agreement with first principles calculations. Finally, we demonstrate the parallel scaling of Allegro with a dynamics simulation of 100 million atoms.