Fitting potential energy surfaces with fundamental invariant neural network. II. Generating fundamental invariants for molecular systems with up to ten atoms

被引:83
作者
Chen, Rongjun [1 ,2 ,3 ]
Shao, Kejie [1 ,2 ]
Fu, Bina [1 ,2 ]
Zhang, Dong H. [1 ,2 ]
机构
[1] Chinese Acad Sci, State Key Lab Mol React Dynam, Dalian Inst Chem Phys, Dalian 116023, Peoples R China
[2] Chinese Acad Sci, Ctr Theoret & Computat Chem, Dalian Inst Chem Phys, Dalian 116023, Peoples R China
[3] Univ Chinese Acad Sci, Beijing 100049, Peoples R China
基金
国家重点研发计划; 中国国家自然科学基金;
关键词
DYNAMICS; REGRESSION;
D O I
10.1063/5.0010104
中图分类号
O64 [物理化学(理论化学)、化学物理学];
学科分类号
070304 ; 081704 ;
摘要
Symmetry adaptation is crucial in representing a permutationally invariant potential energy surface (PES). Due to the rapid increase in computational time with respect to the molecular size, as well as the reliance on the algebra software, the previous neural network (NN) fitting with inputs of fundamental invariants (FIs) has practical limits. Here, we report an improved and efficient generation scheme of FIs based on the computational invariant theory and parallel program, which can be readily used as the input vector of NNs in fitting high-dimensional PESs with permutation symmetry. The newly developed method significantly reduces the evaluation time of FIs, thereby extending the FI-NN method for constructing highly accurate PESs to larger systems beyond five atoms. Because of the minimum size of invariants used in the inputs of the NN, the NN structure can be very flexible for FI-NN, which leads to small fitting errors. The resulting FI-NN PES is much faster on evaluating than the corresponding permutationally invariant polynomial-NN PES.
引用
收藏
页数:8
相关论文
共 49 条
[1]   Gaussian approximation potentials: A brief tutorial introduction [J].
Bartok, Albert P. ;
Csanyi, Gabor .
INTERNATIONAL JOURNAL OF QUANTUM CHEMISTRY, 2015, 115 (16) :1051-1057
[2]   On representing chemical environments (vol 87, 184115, 2013) [J].
Bartok, Albert P. ;
Kondor, Risi ;
Csanyi, Gabor .
PHYSICAL REVIEW B, 2013, 87 (21)
[3]   Generalized neural-network representation of high-dimensional potential-energy surfaces [J].
Behler, Joerg ;
Parrinello, Michele .
PHYSICAL REVIEW LETTERS, 2007, 98 (14)
[4]   First Principles Neural Network Potentials for Reactive Simulations of Large Molecular and Condensed Systems [J].
Behler, Joerg .
ANGEWANDTE CHEMIE-INTERNATIONAL EDITION, 2017, 56 (42) :12828-12840
[5]   Constructing high-dimensional neural network potentials: A tutorial review [J].
Behler, Joerg .
INTERNATIONAL JOURNAL OF QUANTUM CHEMISTRY, 2015, 115 (16) :1032-1050
[6]   Neural network potential-energy surfaces in chemistry: a tool for large-scale simulations [J].
Behler, Joerg .
PHYSICAL CHEMISTRY CHEMICAL PHYSICS, 2011, 13 (40) :17930-17955
[7]   Atom-centered symmetry functions for constructing high-dimensional neural network potentials [J].
Behler, Joerg .
JOURNAL OF CHEMICAL PHYSICS, 2011, 134 (07)
[8]   NEURAL-NETWORK MODELS OF POTENTIAL-ENERGY SURFACES [J].
BLANK, TB ;
BROWN, SD ;
CALHOUN, AW ;
DOREN, DJ .
JOURNAL OF CHEMICAL PHYSICS, 1995, 103 (10) :4129-4137
[9]   The Magma algebra system .1. The user language [J].
Bosma, W ;
Cannon, J ;
Playoust, C .
JOURNAL OF SYMBOLIC COMPUTATION, 1997, 24 (3-4) :235-265
[10]   High-dimensional ab initio potential energy surfaces for reaction dynamics calculations [J].
Bowman, Joel M. ;
Czako, Gabor ;
Fu, Bina .
PHYSICAL CHEMISTRY CHEMICAL PHYSICS, 2011, 13 (18) :8094-8111