MiAD: Mirage Atom Diffusion for De Novo Crystal Generation

TL;DR

MiAD introduces mirage infusion, a technique that allows diffusion models to dynamically alter the number of atoms in crystal generation, significantly improving quality and state-of-the-art S.U.N. rates in de novo crystal design.

Contribution

The paper presents mirage infusion, enabling diffusion models to change atom counts during crystal generation, enhancing variability and performance.

Findings

Model quality improved by up to 2.5x with mirage infusion.

Achieved 8.2% S.U.N. rate on MP-20 dataset.

Outperforms existing state-of-the-art methods.

Abstract

In recent years, diffusion-based models have demonstrated exceptional performance in searching for simultaneously stable, unique, and novel (S.U.N.) crystalline materials. However, most of these models don't have the ability to change the number of atoms in the crystal during the generation process, which limits the variability of model sampling trajectories. In this paper, we demonstrate the severity of this restriction and introduce a simple yet powerful technique, mirage infusion, which enables diffusion models to change the state of the atoms that make up the crystal from existent to non-existent (mirage) and vice versa. We show that this technique improves model quality by up to $\times2.5$ compared to the same model without this modification. The resulting model, Mirage Atom Diffusion (MiAD), is an equivariant joint diffusion model for de novo crystal generation that is capable of altering the number of atoms during the generation process. MiAD achieves an $8.2\%$ S.U.N. rate on the MP-20 dataset, which substantially exceeds existing state-of-the-art approaches. The source code can be found at \href{https://github.com/andrey-okhotin/miad.git}{\texttt{github.com/andrey-okhotin/miad}}.