Neural P$^3$M: A Long-Range Interaction Modeling Enhancer for Geometric GNNs

TL;DR

Neural P$^3$M enhances geometric GNNs by incorporating mesh points and trainable operations, significantly improving long-range interaction modeling and accuracy in molecular energy and force predictions across diverse datasets.

Contribution

This paper introduces Neural P$^3$M, a novel trainable enhancer that extends geometric GNNs to better capture long-range interactions in molecular systems.

Findings

Outperforms benchmarks like MD22 in energy and force prediction.

Achieves 22% average improvement on OE62 dataset.

Compatible with various GNN architectures.

Abstract

Geometric graph neural networks (GNNs) have emerged as powerful tools for modeling molecular geometry. However, they encounter limitations in effectively capturing long-range interactions in large molecular systems. To address this challenge, we introduce Neural P$^3$M, a versatile enhancer of geometric GNNs to expand the scope of their capabilities by incorporating mesh points alongside atoms and reimaging traditional mathematical operations in a trainable manner. Neural P$^3$M exhibits flexibility across a wide range of molecular systems and demonstrates remarkable accuracy in predicting energies and forces, outperforming on benchmarks such as the MD22 dataset. It also achieves an average improvement of 22% on the OE62 dataset while integrating with various architectures.