Topology-Stratified Materials Discovery with A Flow-Based Generative Model
cond-mat.mtrl-sci, cs.AI
Submitted: 2026-09-22
Updated: 2026-09-22
Code: https://github.com/huhuhhhh/UFO-MGen
License: http://creativecommons.org/licenses/by/4.0/
The gist: Accurate generation of crystal structures is the foundation to the discovery of high-performance materials for extreme-environment applications, such as aerospace, additive manufacturing, and fusion
Terminology
Abstract
Accurate generation of crystal structures is the foundation to the discovery of high-performance materials for extreme-environment applications, such as aerospace, additive manufacturing, and fusion energy systems. Although generative modeling has emerged as a promising approach for crystal design, its performance remains limited by the complex crystal structures and diverse chemical compositions. In this work, we develop UFO-MGen, a universal flow-based generative model that learns topological features of Wyckoff representations and leverages this information to accurately generate crystals across vast structural and chemical spaces. Compared with state-of-the-art generative models, UFO-MGen achieves the highest crystal generation success rate under a rigorous multi-stability evaluation framework, the highest SUN (stable, unique, novel) rate, and a remarkable extrapolation capability that has not been reported by previous models. Furthermore, a fine-tuning module is implemented to UFO-MGen for property-constrained crystal generation, enabling the inverse materials design toward target properties. The UFO-MGen opens a new avenue for accelerated materials discovery and providing a foundation for universal materials intelligence.
Sources
- WyckoffDiff -- A Generative Diffusion Model for Crystal Symmetry
- Crystal Structure Prediction by Joint Equivariant Diffusion
- Wyckoff Transformer: Generation of Symmetric Crystals
- SymmCD: Symmetry-Preserving Crystal Generation with Diffusion Models
- Flow Matching for Generative Modeling
- MatterSim: A Deep Learning Atomistic Model Across Elements, Temperatures and Pressures
- Space Group Constrained Crystal Generation
- FlowMM: Generating Materials with Riemannian Flow Matching
- Representing and Learning Functions Invariant Under Crystallographic Groups
- The Geometry of Niggli Reduction III: SAUC -- Search of Alternate Unit Cells
- Towards Symmetry-Aware Generation of Periodic Materials
- $\texttt{Spglib}$: a software library for crystal symmetry search
- Space Group Informed Transformer for Crystalline Materials Generation
Related papers
- Mode-selective electron-phonon coupling drives charge density waves in the kagome metals YRu 3 Si 2 and LaRu 3 Si 2
- Restoring the Surface Magnetic Gap in MnBi 2 Te 4
- ATLAS: A Foundation Neural Sampler for Amorphous Materials
- A Generalizable Machine Learning Framework for Half-Heusler Thermoelectrics
- Nonlinear dynamics in magnonic Fabry-P' e rot resonators: Low-power neuron-like activation and transmission suppression
- AES-Debye: an Accurate, Efficient, and Scalable Engine for Debye Scattering Calculations