Neural Message Passing for Quantum Chemistry

TLDR

Supervised learning on molecular graphs, using symmetry‑invariant neural networks that perform message passing and aggregation, holds great promise for chemistry, drug discovery, and materials science. The study aims to identify an effective variant of this general message‑passing approach and apply it to chemical prediction benchmarks. We unify existing models into a common Message Passing Neural Network framework and investigate novel variations within it. MPNNs achieve state‑of‑the‑art performance on a key molecular property benchmark, suggesting future efforts should target larger molecules or more accurate labels.

Abstract

Supervised learning on molecules has incredible potential to be useful in chemistry, drug discovery, and materials science. Luckily, several promising and closely related neural network models invariant to molecular symmetries have already been described in the literature. These models learn a message passing algorithm and aggregation procedure to compute a function of their entire input graph. At this point, the next step is to find a particularly effective variant of this general approach and apply it to chemical prediction benchmarks until we either solve them or reach the limits of the approach. In this paper, we reformulate existing models into a single common framework we call Message Passing Neural Networks (MPNNs) and explore additional novel variations within this framework. Using MPNNs we demonstrate state of the art results on an important molecular property prediction benchmark; these results are strong enough that we believe future work should focus on datasets with larger molecules or more accurate ground truth labels.