arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2510.11241cond-mat.mtrl-sci
收藏

面向通用机器学习原子间势的跨域迁移优化

Optimizing Cross-Domain Transfer for Universal Machine Learning Interatomic Potentials

Department of Materials Science and Engineering, Seoul National University(首尔大学材料科学与工程系);Research Institute of Advanced Materials, Seoul National University(首尔大学先进材料研究所);Department of Computer Science and Engineering, Seoul National University(首尔大学计算机工程系);AI Center, Samsung Electronics(三星电子人工智能中心);Center for AI and Natural Sciences, Korea Institute of Advanced Study(韩国高级研究院人工智能与自然科学中心)

Jaesun Kim, Jinmu You, Yutack Park, Yunsung Lim, Yujin Kang, Jisu Kim, Haekwan Jeon, Suyeon Ju, Deokgi Hong, Seung Yul Lee, Saerom Choi, Yongdeok Kim, Jae W. Lee, Seungwu Han

AI总结:

本文提出多域训练策略与域桥接集,通过选择性正则化训练 SevenNet-Omni,在跨域精度上达到最先进水平,并实现从 PBE 到 r$^2$SCAN 的有效跨功能迁移。

摘要:

准确且可迁移的机器学习原子间势(MLIPs)对于加速材料与化学发现至关重要。然而,大多数通用 MLIPs 会过拟合于狭窄的数据集或计算协议,限制了它们在化学域和功能域上的可靠性。我们提出了一种可迁移的多域训练策略,通过选择性正则化联合优化通用参数与任务特定参数,并结合域桥接集(DBS)来对齐不同数据集之间的势能面。系统的消融实验表明,较小的 DBS 比例(0.1%)与有针对性的正则化能够协同增强分布外泛化能力,同时保持域内保真度。在涵盖分子、晶体和表面的十五个开放数据库上训练后,我们的模型 SevenNet-Omni 实现了最先进的跨域精度,包括金属表面上的吸附能误差低于 0.06 eV,金属有机框架上的吸附能误差低于 0.1 eV。尽管仅包含 0.5% 的 r$^2$SCAN 数据,SevenNet-Omni 仍能重现高保真的 r$^2$SCAN 能量学,展示了从大型 PBE 数据集进行有效跨功能迁移的能力。该框架为构建通用、可迁移的 MLIPs 提供了一条可扩展的路径,从而桥接量子力学保真度与化学域。

英文摘要:

Accurate yet transferable machine-learning interatomic potentials (MLIPs) are essential for accelerating materials and chemical discovery. However, most universal MLIPs overfit to narrow datasets or computational protocols, limiting their reliability across chemical and functional domains. We introduce a transferable multi-domain training strategy that jointly optimizes universal and task-specific parameters through selective regularization, coupled with a domain-bridging set (DBS) that aligns potential-energy surfaces across datasets. Systematic ablation experiments show that small DBS fractions (0.1%) and targeted regularization synergistically enhance out-of-distribution generalization while preserving in-domain fidelity. Trained on fifteen open databases spanning molecules, crystals, and surfaces, our model, SevenNet-Omni, achieves state-of-the-art cross-domain accuracy, including adsorption-energy errors below 0.06 eV on metallic surfaces and 0.1 eV on metal-organic frameworks. Despite containing only 0.5% r$^2$SCAN data, SevenNet-Omni reproduces high-fidelity r$^2$SCAN energetics, demonstrating effective cross-functional transfer from large PBE datasets. This framework offers a scalable route toward universal, transferable MLIPs that bridge quantum-mechanical fidelities and chemical domains.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑