Orb: A Fast, Scalable Neural Network Potential

TL;DR

Orb is a new family of neural network potentials that are significantly faster and more accurate than existing models, enabling efficient atomistic simulations across diverse materials.

Contribution

The paper introduces Orb, a universal interatomic potential that improves speed and accuracy, and explores foundation model development for materials science.

Findings

Orb models are 3-6 times faster than existing potentials.

Orb achieved a 31% reduction in error on the Matbench Discovery benchmark.

Orb is stable for out-of-distribution materials and effective in various simulation methods.

Abstract

We introduce Orb, a family of universal interatomic potentials for atomistic modelling of materials. Orb models are 3-6 times faster than existing universal potentials, stable under simulation for a range of out of distribution materials and, upon release, represented a 31% reduction in error over other methods on the Matbench Discovery benchmark. We explore several aspects of foundation model development for materials, with a focus on diffusion pretraining. We evaluate Orb as a model for geometry optimization, Monte Carlo and molecular dynamics simulations.