mechanoChemML: A software library for machine learning in computational materials physics
mechanoChemML: A software library for machine learning in computational materials physics
复制标题
DOI：
10.1016/j.commatsci.2022.111493
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
X. Zhang;G. Teichert;Z. Wang;M. Duschenes;S. Srivastava;A. Sunderarajan;E. Livingston;K. Garikipati
中科院分区：
文献类型：
--
作者：
X. Zhang;G. Teichert;Z. Wang;M. Duschenes;S. Srivastava;A. Sunderarajan;E. Livingston;K. Garikipati
我们介绍了mechanoChemML，一个用于计算材料物理的机器学习软件库。mechanoChemML被设计为一方面广泛用于机器学习的平台之间的接口，另一方面用于解决基于偏微分方程的物理模型。这里特别感兴趣的是，mechanoChemML的重点是计算材料物理学的应用。这些典型的特点是耦合的解决方案的材料传输，反应，相变，力学，热传输和电化学。mechanoChemML组织的核心是在数据驱动的计算材料物理学背景下出现的机器学习工作流程。描述了TheMechanoChemML代码结构，布局了机器学习工作流程，并概述了其在解决材料物理学中的几个问题中的应用。
We presentmechanoChemML, a machine learning software library for computational materials physics.mechanoChemMLis designed to function as an interface between platforms that are widely used for machine learning on one hand, and others for solution of partial differential equations-based models of physics. Of special interest here, and the focus ofmechanoChemML, are applications to computational materials physics. These typically feature the coupled solution of material transport, reaction, phase transformation, mechanics, heat transport and electrochemistry. Central to the organization ofmechanoChemMLare machine learning workflows that arise in the context of data-driven computational materials physics. ThemechanoChemMLcode structure is described, the machine learning workflows are laid out and their application to the solution of several problems in materials physics is outlined.