loading page

Improving the performance of fermionic neural networks with the Slater exponential Ansatz
  • +2
  • Denis Bokhan,
  • Maria Kolchenko,
  • Aleksey Boev,
  • Aleksey Fedorov,
  • Dmitrii Trubnikov
Denis Bokhan
Moscow State University

Corresponding Author:denisbokhan@mail.ru

Author Profile
Maria Kolchenko
Moscow State University
Author Profile
Aleksey Boev
Russian Quantum Center
Author Profile
Aleksey Fedorov
Russian Quantum Center
Author Profile
Dmitrii Trubnikov
MSU
Author Profile

Abstract

In this work, we propose a technique for the use of fermionic neural networks (FermiNets) with the Slater exponential Ansatz for electron-nuclear and electron-electron distances, which provides faster convergence of target ground-state energies due to a better description of the interparticle interaction in the vicinities of the coalescence points. Our analysis of learning curves indicates on the possibility to obtain accurate energies with smaller batch sizes using arguments of the bagging approach. In order to obtain even more accurate results for the ground-state energies, we propose an extrapolation scheme for estimating Monte Carlo integrals in the limit of an infinite number of points. Numerical tests for a set of molecules demonstrate a good agreement with the results of the original FermiNets approach (achieved with larger batch sizes than required by our approach) and with results of the coupled-cluster singles and doubles with perturbative triples (CCSD(T)) method that are calculated in the complete basis set (CBS) limit.
24 Oct 2022Review(s) Completed, Editorial Evaluation Pending
24 Oct 2022Submitted to International Journal of Quantum Chemistry
26 Oct 2022Assigned to Editor
26 Oct 2022Submission Checks Completed
27 Oct 2022Reviewer(s) Assigned
31 Oct 2022Reviewer(s) Assigned