Multimodal Crystal Flow: Any-to-Any Modality Generation for Unified Crystal Modeling

TL;DR

This paper introduces MCFlow, a unified multimodal flow model that enables various crystal generation tasks within a single framework by leveraging a novel atom ordering and permutation augmentation, achieving competitive results.

Contribution

The paper presents MCFlow, a novel unified multimodal flow model for crystal generation that incorporates symmetry-aware atom ordering and hierarchical augmentation.

Findings

Achieves competitive performance on MP-20 and MPTS-52 benchmarks.

Unifies multiple crystal generation tasks within a single model.

Incorporates symmetry and compositional priors without explicit templates.

Abstract

Crystal modeling spans a family of conditional and unconditional generation tasks across different modalities, including crystal structure prediction (CSP) and \emph{de novo} generation (DNG). While recent deep generative models have shown promising performance, they remain largely task-specific, lacking a unified framework that shares crystal representations across different generation tasks. To address this limitation, we propose \emph{Multimodal Crystal Flow (MCFlow)}, a unified multimodal flow model that realizes multiple crystal generation tasks as distinct inference trajectories via independent time variables for atom types and crystal structures. To enable multimodal flow in a standard transformer model, we introduce a composition- and symmetry-aware atom ordering with hierarchical permutation augmentation, injecting strong compositional and crystallographic priors without explicit structural templates. Experiments on the MP-20 and MPTS-52 benchmarks show that MCFlow achieves competitive performance against task-specific baselines across multiple crystal generation tasks.