.png){kind=icon}
返回
Data-Efficient Multifidelity Training for High-Fidelity Machine Learning Interatomic Potentials
.png){kind=icon}
PRE
AI
DOI：10.1021/jacs.4c14455.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Machine learning interatomic potentials (MLIPs) are used to estimate potential energy surfaces (PES) from ab initio calculations, providing near-quantum-level accuracy with reduced computational costs. However, the high cost of assembling high-fidelity databases hampers the application of MLIPs to systems that require high chemical accuracy. Utilizing an equivariant graph neural network, we present an MLIP framework that trains on multifidelity databases simultaneously. This approach enables the accurate learning of high-fidelity PES with minimal high-fidelity data. Employing the generalized gradient approximation (GGA) and meta-GGA as low- and high-fidelity approaches, respectively, we tested this framework on the Li6PS5Cl and In x Ga1-x N systems. The results show that using a high-fidelity training set with a size approximately 10% of the low-fidelity set, the multifidelity training framework achieves excellent accuracy, with Li-ion conductivity predictions within 10% error and In x Ga1-x N mixing energy showing an R 2 of 0.98 compared to the reference high-fidelity MLIP results. It indicates that geometric and compositional spaces not covered by the high-fidelity meta-GGA database can be effectively inferred from low-fidelity GGA data, thus enhancing accuracy and molecular dynamics stability. We also developed a general-purpose MLIP that utilizes both GGA and meta-GGA data from the Materials Project, significantly enhancing MLIP performance for high-accuracy tasks such as predicting energies above hull for crystals in general. Furthermore, we demonstrate that the present multifidelity learning is more effective than transfer learning or Delta-learning and that it can also be applied to learn higher-fidelity up to the coupled-cluster level. We believe this methodology holds promise for creating highly accurate bespoke or universal MLIPs by effectively expanding the high-fidelity data set.
Keyword:
DENSITY-FUNCTIONAL THEORY
期刊
IF:
15.6
论文数:
20.0W
被引数:
60.2W
机构
引用论文
Job tenure and quality of work life of people with psychiatric disabilities working in social enterprises在社会企业工作的精神病患者的工作任期和工作生活质量
PREAI
分享收藏E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentialsE(3)-等变图神经网络，用于数据高效和精确的原子间势.png)
NATURE COMMUNICATIONS
OAAI
NATURE COMMUNICATIONSIF15.7
OAAI分享收藏Learning together: Towards foundation models for machine learning interatomic potentials with meta-learning共同学习: 通过元学习建立机器学习的原子间势的基础模型
OAAI
OAAI分享收藏Approaching coupled cluster accuracy with a general-purpose neural network potential through transfer learning通过迁移学习接近具有通用神经网络潜力的耦合聚类精度NATURE COMMUNICATIONSOAAI
NATURE COMMUNICATIONSIF15.7
OAAI分享收藏PubChemQC Project: A Large-Scale First-Principles Electronic Structure Database for Data-Driven ChemistryPubChemQC项目: 用于数据驱动化学的大型第一原理电子结构数据库
PREAI
分享收藏Transfer Learning for Affordable and High-Quality Tunneling Splittings from Instanton CalculationsOAAI
OAAI分享收藏CHGNet as a pretrained universal neural network potential for charge-informed atomistic modellingCHGNet作为预训练的通用神经网络，具有用于电荷信息原子建模的潜力OAAI
OAAI分享收藏
查看更多内容