TorchMD-NET: Equivariant Transformers for Neural Network based Molecular Potentials

TL;DR

TorchMD-NET introduces an equivariant transformer architecture that significantly improves the accuracy and efficiency of neural network-based molecular potential predictions, providing new insights into molecular conformations and emphasizing the importance of diverse datasets.

Contribution

The paper presents TorchMD-NET, a novel equivariant transformer architecture that outperforms existing models in molecular property prediction tasks.

Findings

Outperforms state-of-the-art on MD17, ANI-1, and QM9 datasets.

Provides insights into learned molecular representations.

Highlights the importance of including off-equilibrium conformations in datasets.

Abstract

The prediction of quantum mechanical properties is historically plagued by a trade-off between accuracy and speed. Machine learning potentials have previously shown great success in this domain, reaching increasingly better accuracy while maintaining computational efficiency comparable with classical force fields. In this work we propose TorchMD-NET, a novel equivariant transformer (ET) architecture, outperforming state-of-the-art on MD17, ANI-1, and many QM9 targets in both accuracy and computational efficiency. Through an extensive attention weight analysis, we gain valuable insights into the black box predictor and show differences in the learned representation of conformers versus conformations sampled from molecular dynamics or normal modes. Furthermore, we highlight the importance of datasets including off-equilibrium conformations for the evaluation of molecular potentials.