Study on potential energy curves and ro-vibrational energies of DT, HT and T2 molecules

Spectroscopic constants from atomic properties: a machine learning approach

“…9,41,47 More recently, Fu et al used an ML model to predict dissociation energies for diatomic molecules, exploiting microscopic and macroscopic properties. 150 They tested their model against CO and highlighted that the reported experimental dissociation energy in the literature had increased by 100 kcal mol −1 over the course of 78 years from 1936 to 2014 (ref. [150][151][152] (in Table 1 of ref.…”

“…150 They tested their model against CO and highlighted that the reported experimental dissociation energy in the literature had increased by 100 kcal mol −1 over the course of 78 years from 1936 to 2014 (ref. [150][151][152] (in Table 1 of ref. 150).…”

“…[150][151][152] (in Table 1 of ref. 150). The data used to train model d1 is primarily collected from Huber and Herzberg's constants of diatomic molecules, rst published in 1979.…”

Combining ab initio and machine learning method to improve the prediction of diatomic vibrational energies

“…According to our knowledge, there is no effective method to uniformly pre evaluate the prediction performance of various ab initio calculation configurations, so the calculation for real system needs to go through a cumbersome trial and error process. For the calculation of vibrational levels of diatomic molecules, the uncertainty of ab initio results can be hundreds or even thousands cm À1 [12,17,18]. In order to improve prediction performance, traditional data-driven algorithm constructs a specific parameter model for a specific system, and extracts detail parameter information from experimental data of the research object.…”

Achieving vibrational energies of diatomic systems with high quality by machine learning improved DFT method

“…[9][10][11][12][13] The post-Hartree Fock methods like multireference conguration interaction methods have decent performance in accuracy, whereas the steep computation cost and lengthy time make them limited to small systems. 12,14,15 The DFT method is compromised in accuracy, but it has rapid calculation speed, which makes it the rst choice for the calculation of large systems. [16][17][18] In order to improve the performance of DFT, both general and accurate exchange-correlation functionals and basis set are required, 16,[18][19][20] which is still a challenge to us.…”