MatMind: A Structure-Activity Knowledge-Driven Generative Foundation Model for Materials Science
On the map: Diffusion models → Flow matching, time series and science
Abstract (arXiv)
Progress in AI-driven crystal materials science has so far been carried by narrow architectures purpose-built for individual tasks -- graph neural networks for property prediction, diffusion and flow-matching models for crystal generation -- each excelling within its niche yet unable to act as a shared backbone across the full spectrum of materials problems. Generative large language models offer a fundamentally different paradigm, in which structural representation, quantitative prediction, and structure-activity reasoning can be unified within one model, but the materials community has yet to see this paradigm realized at a level competitive with established narrow specialists. Here we present MatMind, a generative foundation model purpose-built for crystal materials science under this paradigm, developed through the coordinated activation of structure-activity knowledge and physics-informed feedback within a progressive training framework -- combining structure-activity knowledge injection, a dual-head architecture that jointly trains language reasoning and numerical regression in a shared representation space, and multi-objective physics-informed reinforcement learning over stability, novelty, and structural diversity. Across three task families, MatMind attains the lowest mean absolute error on energy above hull, bulk modulus, and band gap -- surpassing graph neural network predictors purpose-built for these tasks -- reaches an S.U.N. rate of 65.3% on unconditional crystal generation, and achieves a comparable multiplicative improvement on magnetization-density-conditioned generation, where only 21 positive samples exist within over 600000 training entries. By matching or surpassing narrow specialists on their own ground while operating within a single unified model, MatMind shows that the LLM-based paradigm can serve as a viable backbone for crystal materials science going forward.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 28 of its 82 references that are in the map. First posted on arXiv 5 Jun 2026.
Builds on
28 of 82 references.
Closest ancestors · 1
Same direction, newest first
Other references · 27
- 2022Training language models to follow instructions with human feedbackcited by 15,712
- 2017Proximal Policy Optimization Algorithmscited by 14,441
- 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,302
- 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
- 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
- 2019SciBERT: A Pretrained Language Model for Scientific Textcited by 1,404
- 2023Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizescited by 710
- 2021E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentialscited by 701
- 2021Crystal Diffusion Variational Autoencoder for Periodic Material Generationcited by 268
- 2017Deep Reinforcement Learning for De-Novo Drug Designcited by 161
- 2019Do NLP Models Know Numbers? Probing Numeracy in Embeddingscited by 157
- 2021Delving into Deep Imbalanced Regressioncited by 157
15 more
- 2020Injecting Numerical Reasoning Skills into Language Modelscited by 133
- 2023Language models can generate molecules, materials, and protein binding sites directly in three dimensions as XYZ, CIF, and PDB filescited by 59
- 2024Open Materials 2024 (OMat24) Inorganic Materials Dataset and Modelscited by 50
- 2024Space Group Informed Transformer for Crystalline Materials Generationcited by 40
- 2023Matbench Discovery -- A framework to evaluate machine learning crystal stability predictionscited by 34
- 2025LeMat-GenBench: A Unified Evaluation Framework for Crystal Generative Modelscited by 20
- 2025Guiding Generative Models to Uncover Diverse and Novel Crystals via Reinforcement Learningcited by 19
- 2022Deep Reinforcement Learning for Inverse Inorganic Materials Designcited by 18
- 2023CrysMMNet: Multimodal Representation for Crystal Property Predictioncited by 18
- 2024Foundational Large Language Models for Materials Researchcited by 18
- 2024Structure-based out-of-distribution (OOD) materials property prediction: a benchmark studycited by 17
- 2025Reinforcement Fine-Tuning for Materials Designcited by 11
- 2021Audacity of huge: overcoming challenges of data scarcity and data quality for machine learning in computational materials discoverycited by 10
- 2025Siamese Foundation Models for Crystal Structure Predictioncited by 8
- 2026CrystalReasoner: Reasoning and RL for Property-Conditioned Crystal Structure Generationcited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2025AtomWorld: A Benchmark for Evaluating Spatial Reasoning in Large Language Models on Crystalline Materials
- 2024Generative Hierarchical Materials Search
- 2026A collaborative agent with two lightweight synergistic models for autonomous crystal materials research
- 2026Reinforcement Learning on the Discrete Composition Channel of a Crystal Generator: Validated Gains and Reward Hacking
- 2025A Survey of AI for Materials Science: Foundation Models, LLM Agents, Datasets, and Tools
- 2025PLaID++: A Preference Aligned Language Model for Targeted Inorganic Materials Design
- 2026SCALAR: Quantifying Structural Hallucination, Consistency, and Reasoning Gaps in Materials Foundation Models
- 2026Mat-Pref: Verifiable-Reward Training Improves Compositional Reasoning in Inorganic Materials