Deep Potential: a general representation of a many-body potential energy surface

TL;DR

Deep Potential introduces a neural network-based method for accurately modeling potential energy surfaces in atomic and molecular systems, combining first-principles accuracy with computational efficiency and scalability.

Contribution

It provides a general, symmetry-respecting deep learning framework for potential energy surfaces that does not rely on empirical fitting or approximations.

Findings

Reproduces quantum mechanics accuracy within chemical precision

Computational cost comparable to empirical force fields

Scalable to large systems

Abstract

We present a simple, yet general, end-to-end deep neural network representation of the potential energy surface for atomic and molecular systems. This methodology, which we call Deep Potential, is "first-principle" based, in the sense that no ad hoc approximations or empirical fitting functions are required. The neural network structure naturally respects the underlying symmetries of the systems. When tested on a wide variety of examples, Deep Potential is able to reproduce the original model, whether empirical or quantum mechanics based, within chemical accuracy. The computational cost of this new model is not substantially larger than that of empirical force fields. In addition, the method has promising scalability properties. This brings us one step closer to being able to carry out molecular simulations with accuracy comparable to that of quantum mechanics models and computational cost comparable to that of empirical potentials.