Factor Graph Molecule Network for Structure Elucidation

TL;DR

This paper introduces a novel factor graph neural network that effectively learns molecular structures from physical and chemical properties, improving generalization and enforcing chemical constraints for drug discovery.

Contribution

It combines higher-order relational learning with neural networks to enhance molecule structure prediction, addressing efficiency, parameter sharing, and symmetry challenges.

Findings

Outperforms related methods in structure prediction accuracy

Effective in enforcing valence and higher-order relationships

Demonstrates strong generalization in molecular learning tasks

Abstract

Designing a network to learn a molecule structure given its physical/chemical properties is a hard problem, but is useful for drug discovery tasks. In this paper, we incorporate higher-order relational learning of Factor Graphs with strong approximation power of Neural Networks to create a molecule-structure learning network that has strong generalization power and can enforce higher-order relationship and valence constraints. We further propose methods to tackle problems such as the efficient design of factor nodes, conditional parameter sharing among factors, and symmetry problems in molecule structure prediction. Our experiment evaluation shows that the factor learning is effective and outperforms related methods.