# Adversarial Learned Molecular Graph Inference and Generation

## TL;DR

This paper introduces ALMGIG, a novel adversarial learning framework for molecular graph inference and generation that avoids expensive graph isomorphism computations and effectively captures molecular properties for drug discovery.

## Contribution

ALMGIG extends GANs with cycle-consistency and a multi-graph GIN to improve molecular distribution learning and generation without explicit reconstruction loss.

## Key findings

- ALMGIG outperforms baselines in learning molecular distributions.
- It enables efficient molecule search in latent space for drug discovery.
- The method accurately captures physicochemical property distributions.

## Abstract

Recent methods for generating novel molecules use graph representations of molecules and employ various forms of graph convolutional neural networks for inference. However, training requires solving an expensive graph isomorphism problem, which previous approaches do not address or solve only approximately. In this work, we propose ALMGIG, a likelihood-free adversarial learning framework for inference and de novo molecule generation that avoids explicitly computing a reconstruction loss. Our approach extends generative adversarial networks by including an adversarial cycle-consistency loss to implicitly enforce the reconstruction property. To capture properties unique to molecules, such as valence, we extend the Graph Isomorphism Network to multi-graphs. To quantify the performance of models, we propose to compute the distance between distributions of physicochemical properties with the 1-Wasserstein distance. We demonstrate that ALMGIG more accurately learns the distribution over the space of molecules than all baselines. Moreover, it can be utilized for drug discovery by efficiently searching the space of molecules using molecules' continuous latent representation. Our code is available at https://github.com/ai-med/almgig