MHG-GNN: Combination of Molecular Hypergraph Grammar with Graph Neural Network

TL;DR

This paper introduces MHG-GNN, an autoencoder that combines molecular hypergraph grammar with graph neural networks to improve property prediction in material science, aiming to develop a foundation model for the field.

Contribution

The paper presents a novel autoencoder integrating molecular hypergraph grammar with GNNs, advancing material property prediction methods.

Findings

MHG-GNN shows promising results across diverse property prediction tasks.

The approach effectively captures complex molecular structures.

Results indicate potential for foundation models in material science.

Abstract

Property prediction plays an important role in material discovery. As an initial step to eventually develop a foundation model for material science, we introduce a new autoencoder called the MHG-GNN, which combines graph neural network (GNN) with Molecular Hypergraph Grammar (MHG). Results on a variety of property prediction tasks with diverse materials show that MHG-GNN is promising.