期刊
JOURNAL OF CHEMICAL PHYSICS
卷 148, 期 24, 页码 -出版社
AIP Publishing
DOI: 10.1063/1.5011181
关键词
-
资金
- Laboratory Directed Research and Development (LDRD) program
- Advanced Simulation and Computing (ASC) program
- Center for Nonlinear Studies (CNLS) at Los Alamos National Laboratory (LANL)
We introduce the Hierarchically Interacting Particle Neural Network (HIP-NN) to model molecular properties from datasets of quantum calculations. Inspired by a many-body expansion, HIP-NN decomposes properties, such as energy, as a sum over hierarchical terms. These terms are generated from a neural network-a composition of many nonlinear transformations-acting on a representation of the molecule. HIP-NN achieves the state-of-the-art perfo(r)mance on a dataset of 131k ground state organic molecules and predicts energies with 0.26 kcal/mol mean absolute error. With minimal tuning, our model is also competitive on a dataset of molecular dynamics trajectories. In addition to enabling accurate energy predictions, the hierarchical structure of HIP-NN helps to identify regions of model uncertainty. Published by AIP Publishing.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据