用物理先验提升晶体生成模型的稳定性，无需额外计算开销。
CrystalREPA: Transferring Physical Priors from Universal MLIPs to Crystal Generative Models
- 通过对比学习对齐生成模型与通用势函数的原子级表征。
- 在多个数据集上显著提升生成晶体的热力学稳定性和结构保真度。
- 提出新选师标准：看原子表示空间可区分性，而非传统精度排行榜。
晶体生成模型主要学习稳定晶体的外观，但缺乏对稳定性的显式监督。我们通过能量探测发现，当前先进晶体生成模型与预训练通用机器学习原子势函数（MLIPs）之间存在显著表征差异，并证明可通过简单的训练时对齐来弥合这一差距。本文提出Crystal REPresentation Alignment（CrystalREPA），一个即插即用框架，利用元素感知的对比目标，将生成编码器的原子级隐藏状态与冻结的MLIP表示对齐，以极小训练成本转移具备稳定性意识的原子先验，且不增加推理开销。在三种生成框架、十种MLIP教师和两个基准数据集上，CrystalREPA一致提升了生成晶体的热力学稳定性、结构有效性与结构保真度。更重要的是，我们发现MLIP的迁移效果难以由标准排行榜（如Matbench Discovery）的精度预测，却强烈依赖于其原子级表示空间的可区分性，从而提供了一种独立于精度的实用教师选择准则。
原文摘要 · Abstract (English)
Crystal generative models mainly learn what stable crystals look like, with little explicit supervision for what makes them stable. We reveal a substantial representation gap between state-of-the-art crystal generative models and pretrained universal machine learning interatomic potentials (MLIPs) via energy probing, and show this gap can be closed by a simple training-time alignment. We propose Crystal REPresentation Alignment (CrystalREPA), a plug-and-play framework that aligns the atom-wise hidden states of generative encoders with frozen MLIP representations through an element-aware contrastive objective, transferring stability-aware atomistic priors with marginal training overhead and no additional inference cost. Across three generative frameworks, ten MLIP teachers, and two benchmark datasets, CrystalREPA consistently improves the thermodynamic stability, structural validity, and structural fidelity of generated crystals. Equally important, we find that an MLIP's transfer effectiveness is poorly predicted by its accuracy on standard leaderboards (e.g., Matbench Discovery) but strongly predicted by the distinguishability of its atom-wise representation space, yielding a practical, accuracy-independent criterion for selecting MLIP teachers for generative transfer.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。