Text-guided flow matching enables sample-efficient crystal structure generation
On the map: Diffusion models → Flow matching, time series and science
Abstract (arXiv)
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.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 19 of its 53 references that are in the map. First posted on arXiv 1 Sep 2026.
Builds on
19 of 53 references.
Closest ancestors · 6
Same direction, newest first
- 2025Periodic Materials Generation using Text-Guided Joint Diffusion Model1 Mar
- 2024FlowLLM: Flow Matching for Material Generation with Large Language Models as Base Distributions30 Oct
- 2023Crystal Structure Prediction by Joint Equivariant Diffusion30 Jul
- 2023Improving and generalizing flow-based generative models with minibatch optimal transport1 Feb
- 2022Flow Matching for Generative Modeling6 Oct
- 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow7 Sep
Other references · 13
- 2020Denoising Diffusion Probabilistic Modelscited by 21,076
- 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
- 2022Classifier-Free Diffusion Guidancecited by 5,847
- 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 3,045
- 2019SciBERT: A Pretrained Language Model for Scientific Textcited by 1,414
- 2022Building Normalizing Flows with Stochastic Interpolantscited by 995
- 2021Crystal Diffusion Variational Autoencoder for Periodic Material Generationcited by 270
- 2024FlowMM: Generating Materials with Riemannian Flow Matchingcited by 109
- 2021Inverse design of 3d molecular structures with conditional generative neural networkscited by 93
- 2023MatSci-NLP: Evaluating Scientific Language Models on Materials Science Language Tasks Using Text-to-Schema Modelingcited by 35
- 2022Physics Guided Deep Learning for Generative Design of Crystal Materials with Symmetry Constraintscited by 30
- 2023Towards Symmetry-Aware Generation of Periodic Materialscited by 26
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Lang2Str: Two-Stage Crystal Structure Generation with LLMs and Continuous Flow Models
- 2026DMFlow: Disordered Materials Generation by Flow Matching
- 2026uFlowCSP: Crystal Structure Prediction using Mean flow generative models
- 2025LLM Meets Diffusion: A Hybrid Framework for Crystal Material Generation
- 2026Property-Informed Diffusion-Based Text-to-Microstructure Generation
- 2024Foundational Large Language Models for Materials Research
- 2026EP-Flow: Disordered Crystal Structure Prediction without Site-Level Annotations
- 2026Composable Crystals: Controllable Materials Discovery via Concept Learning