Inverse design of 3d molecular structures with conditional generative neural networks

TL;DR

This paper introduces a conditional generative neural network for designing 3D molecular structures with specific properties, enabling targeted and novel molecule generation even with limited reference data.

Contribution

It presents a new neural network approach that is agnostic to chemical bonding, allowing for inverse design of molecules with desired features and properties.

Findings

Successfully generated molecules with specified motifs and compositions

Discovered stable molecules with targeted electronic properties

Enabled joint targeting of multiple properties beyond training data

Abstract

The rational design of molecules with desired properties is a long-standing challenge in chemistry. Generative neural networks have emerged as a powerful approach to sample novel molecules from a learned distribution. Here, we propose a conditional generative neural network for 3d molecular structures with specified chemical and structural properties. This approach is agnostic to chemical bonding and enables targeted sampling of novel molecules from conditional distributions, even in domains where reference calculations are sparse. We demonstrate the utility of our method for inverse design by generating molecules with specified motifs or composition, discovering particularly stable molecules, and jointly targeting multiple electronic properties beyond the training regime.