Permutation-Invariant-Polynomial Neural-Network-Based Delta-Machine Learning Approach: A Case for the HO2 Self-Reaction and Its Dynamics Study

D-machine learning, or the hierarchical construction scheme, is a highly cost-effective method, as only a small number of high-level ab initio energies are required to improve a potential energy surface (PES) fit to a large number of low-level points. However, there is no efficient and systematic way to select as few points as possible from the low-level data set. We here propose a permutation-invariant-polynomial neural-network (PIP-NN)based Delta-machine learning approach to construct full-dimensional accurate PESs of complicated reactions efficiently. Particularly, the high flexibility of the NN is exploited to efficiently sample points from the low-level data set. This approach is applied to the challenging case of a HO2 self-reaction with a large configuration space. Only 14% of the DFT data set is used to successfully bring a newly fitted DFT PES to the UCCSD(T)-F12a/AVTZ quality. Then, the quasiclassical trajectory (QCT) calculations are performed to study its dynamics, particularly the mode specificity.

Automated summary

The study introduces a neural network-based Delta-machine learning approach for efficiently constructing full-dimensional accurate potential energy surfaces of complex reactions. The flexibility of the neural network is utilized to efficiently sample points from the low-level data set and successfully elevate the newly fitted potential energy surface to a high-quality level.