Modular Flows: Differential Molecular Generation

TL;DR

This paper introduces a novel class of E(3)-equivariant continuous normalizing flows for molecular generation, addressing limitations of previous models by ensuring permutation invariance and validity without post-processing, achieving state-of-the-art results.

Contribution

The authors develop continuous normalizing flows based on graph PDEs that improve molecular generation by ensuring invariance and validity, outperforming existing methods.

Findings

Achieves state-of-the-art performance on QM9 and ZINC250K benchmarks.

Models are cast as message-passing temporal networks.

Effectively addresses issues of dequantization and permutation invariance.

Abstract

Generating new molecules is fundamental to advancing critical applications such as drug discovery and material synthesis. Flows can generate molecules effectively by inverting the encoding process, however, existing flow models either require artifactual dequantization or specific node/edge orderings, lack desiderata such as permutation invariance, or induce discrepancy between the encoding and the decoding steps that necessitates post hoc validity correction. We circumvent these issues with novel continuous normalizing E(3)-equivariant flows, based on a system of node ODEs coupled as a graph PDE, that repeatedly reconcile locally toward globally aligned densities. Our models can be cast as message-passing temporal networks, and result in superlative performance on the tasks of density estimation and molecular generation. In particular, our generated samples achieve state-of-the-art on both the standard QM9 and ZINC250K benchmarks.