Using an Autoencoder for Dimensionality Reduction in Quantum Dynamics

被引:2
作者
Reiter, Sebastian [1 ]
Schnappinger, Thomas [1 ]
de Vivie-Riedle, Regina [1 ]
机构
[1] Ludwig Maximilians Univ Munchen, Dept Chem, Butenandtstr 5-13, Munich, Germany
来源
ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING - ICANN 2019: WORKSHOP AND SPECIAL SESSIONS | 2019年 / 11731卷
关键词
Quantum dynamics; Machine learning; Autoencoder; Dimensionality reduction;
D O I
10.1007/978-3-030-30493-5_73
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
A key step in performing quantum dynamics for a chemical system is the reduction of dimensionality to allow a numerical treatment. Here, we introduce a machine learning approach for the (semi)automatic construction of reactive coordinates. After generating a meaningful data set from trajectory calculations, we train an autoencoder to find a low-dimensional set of non-linear coordinates for use in molecular quantum dynamics. We compare the wave packet dynamics of proton transfer reactions in both linear and non-linear coordinate spaces and find significant improvement for physical properties like reaction timescales.
引用
收藏
页码:783 / 787
页数:5
相关论文
共 6 条