Conditional $\beta$-VAE for De Novo Molecular Generation

TL;DR

This paper introduces a recurrent, conditional $eta$-VAE that disentangles the latent space to improve de novo molecular generation, property optimization, and validity, achieving state-of-the-art results on key chemical metrics.

Contribution

The paper presents a novel recurrent, conditional $eta$-VAE with a mutual information driven training protocol that enhances molecule validity and optimization capabilities.

Findings

Achieves SOTA on penalized LogP (pLogP) optimization.

Matches SOTA for validity, novelty, and uniqueness in molecule generation.

Demonstrates improved constrained optimization results.

Abstract

Deep learning has significantly advanced and accelerated de novo molecular generation. Generative networks, namely Variational Autoencoders (VAEs) can not only randomly generate new molecules, but also alter molecular structures to optimize specific chemical properties which are pivotal for drug-discovery. While VAEs have been proposed and researched in the past for pharmaceutical applications, they possess deficiencies which limit their ability to both optimize properties and decode syntactically valid molecules. We present a recurrent, conditional $\beta$-VAE which disentangles the latent space to enhance post hoc molecule optimization. We create a mutual information driven training protocol and data augmentations to both increase molecular validity and promote longer sequence generation. We demonstrate the efficacy of our framework on the ZINC-250k dataset, achieving SOTA unconstrained optimization results on the penalized LogP (pLogP) and QED scores, while also matching current SOTA results for validity, novelty and uniqueness scores for random generation. We match the current SOTA on QED for top-3 molecules at 0.948, while setting a new SOTA for pLogP optimization at 104.29, 90.12, 69.68 and demonstrating improved results on the constrained optimization task.