Fast and Uncertainty-Aware Directional Message Passing for Non-Equilibrium Molecules

TL;DR

This paper introduces DimeNet++, a faster and more accurate message passing model for predicting properties of molecules during reactions, including non-equilibrium states, with uncertainty quantification methods.

Contribution

The paper presents DimeNet++, an improved model for non-equilibrium molecules, and introduces the COLL dataset for reactive molecules, advancing machine learning in chemical reaction modeling.

Findings

DimeNet++ is 8x faster and 10% more accurate than DimeNet on QM9.

DimeNet++ performs well on the COLL dataset of reactive molecules.

Uncertainty quantification methods help accelerate exploration of non-equilibrium structures.

Abstract

Many important tasks in chemistry revolve around molecules during reactions. This requires predictions far from the equilibrium, while most recent work in machine learning for molecules has been focused on equilibrium or near-equilibrium states. In this paper we aim to extend this scope in three ways. First, we propose the DimeNet++ model, which is 8x faster and 10% more accurate than the original DimeNet on the QM9 benchmark of equilibrium molecules. Second, we validate DimeNet++ on highly reactive molecules by developing the challenging COLL dataset, which contains distorted configurations of small molecules during collisions. Finally, we investigate ensembling and mean-variance estimation for uncertainty quantification with the goal of accelerating the exploration of the vast space of non-equilibrium structures. Our DimeNet++ implementation as well as the COLL dataset are available online.