Hyperparameter optimization in machine learning algorithm for extrapolations of variation calculations
An optimization method for hyperparameters of our machine learning extrapolation algorithm for results of variational calculations in quantum mechanics, is proposed. The method makes it possible to obtain the optimal values of hyperparameters for artificial neural network training. Deviations of some hyperparameters from the optimal values are shown to result in distorting the extrapolation predictions.
Download file
Counter downloads: 49
Keywords
machine learning, extrapolation methods, bound state energies, nuclear shell modelAuthors
| Name | Organization | |
| Belozerov A.O. | Pacific National University | aobelozerov@gmail.com |
| Mazur A.I. | Pacific National University | mazur@khb.ru |
| Shirokov A.M. | Skobeltsyn Institute of Nuclear Physics M.V. Lomonosov Moscow State University | shirokov@nucl-th.sinp.msu.ru |
References
Barrett B.R., Navrátil P., Vary J.P. // Prog. Part. Nucl. Phys. - 2013. - V. 69(1). - P. 131-181.
Maris P., Vary J.P., Shirokov A.M. // Phys. Rev. C. - 2009. - V. 79. - P. 014308.
Athanassopoulos S., Mavrommatis E., Gernoth K., Clark J. // Nucl. Phys. A. - 2004. - V. 743. - P. 222.
Akkoyun S., Bayram T., Kara S.O., Sinan A. //j. Phys. G. - 2013. - V. 40. - P. 055106.
Athanassopoulos S., Mavrommatis E., Gernoth K.A., Clark J.W. // arXiv:0509075 [nuclth].
David C., Freslier M., Aichelin J. // Phys. Rev. C. - 1995. - V. 51. - P. 1453.
Bass S.A., Bischoff A., Maruhn J.A., et al. // Phys. Rev. C. - 1996. - V. 53. - P. 2358.
Haddad F. et al. // Phys. Rev. C. - 1997. - V. 55. - P. 1371.
Costiris N., Mavrommatis E., Gernoth K.A., Clark J.W. // arXiv:0701096 [nucl-th].
Shirokov A.M., Shin I.J., Kim Y., et al. // Phys. Lett. B. - 2016. - V. 761. - P. 87-91.
Negoita G.A., Vary J.P., Luecke G.R., et al. // Phys. Rev. C. - 2019. - V. 99. - P. 054308.
Kingma D., Ba J. // arXiv:1412.6980 [cs.LG].
Probst P., Boulesteix A., Bischl B. //j. Mach. Learn. Res. - 2019. - V. 20. - P. 1-32.
