Text-guided flow matching enables sample-efficient crystal structure generation
Abstract
Crystal generators can now propose periodic structures, but their control interfaces remain poorly matched to the mixed descriptors used in materials design. Text provides a compact way to combine composition, symmetry, prototype and property cues, yet it has not been clear whether such information can steer flow-based crystal generation. Here we introduce TFMat, a text-conditioned flow-matching framework that uses structured materials language as a semantic prior for a CrystalFlow generator. Across Perov-5, Carbon-24 and MP-20 crystal structure prediction benchmarks, TFMat improves one-candidate match rates over CrystalFlow and reaches a 92.04% MP-20 match rate with 20 candidates; in de novo generation, it improves element-count and density distribution alignment while retaining coarse property consistency in composition-selected outputs. These results position structured text as an inspectable control layer for translating human-readable materials intent into candidate crystals for downstream simulation and validation.
Related
- SymADiT: Symmetry-Aware Crystal Generation with Latent Flow Matching
- Phase matching and mode coupling in photonic-crystal ring resonators
- Automated mode-aware design of nonlinear photonic crystal cavities for intrinsic second harmonic generation
- 266 nm generation performance of NSBBF crystal improved by elevating the operating temperature
- Crystal orientation enables near-unity nonlinear circular dichroism in resonant metasurfaces
BibTeX
@misc{li2026textguided,
title = {{Text-guided flow matching enables sample-efficient crystal structure generation}},
author = {Wentao Li},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.01076},
eprint = {2609.01076},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.01076},
note = {CC BY-NC-SA 4.0},
}
RIS
TY - UNPB
TI - Text-guided flow matching enables sample-efficient crystal structure generation
AU - Wentao Li
PY - 2026
DA - 2026/09/06
PB - arXiv
DO - 10.48550/arxiv.2609.01076
UR - https://arxiv.org/abs/2609.01076
AB - Crystal generators can now propose periodic structures, but their control interfaces remain poorly matched to the mixed descriptors used in materials design. Text provides a compact way to combine composition, symmetry, prototype and property cues, yet it has not been clear whether such information can steer flow-based crystal generation. Here we introduce TFMat, a text-conditioned flow-matching framework that uses structured materials language as a semantic prior for a CrystalFlow generator. Across Perov-5, Carbon-24 and MP-20 crystal structure prediction benchmarks, TFMat improves one-candidate match rates over CrystalFlow and reaches a 92.04% MP-20 match rate with 20 candidates; in de novo generation, it improves element-count and density distribution alignment while retaining coarse property consistency in composition-selected outputs. These results position structured text as an inspectable control layer for translating human-readable materials intent into candidate crystals for downstream simulation and validation.
ER -