Machine Learning Modeling of Wigner Intracule Functionals for Two Electrons in One Dimension

TL;DR

This paper uses machine learning to model a universal functional transformation for two-electron systems in one dimension, achieving accurate correlation energy predictions from Wigner distribution functions.

Contribution

It introduces a kernel-based machine learning approach to approximate the functional transformation in Wigner distribution theory for two-electron systems, with a novel regularization method.

Findings

Achieved sub-chemical accuracy in correlation energy predictions.

Demonstrated effective modeling using Hartree-Fock level Wigner functions.

Provided a dataset of 923 one-dimensional potentials for training and testing.

Abstract

In principle, many-electron correlation energy can be precisely computed from a reduced Wigner distribution function ($\mathcal{W}$) thanks to a universal functional transformation ($\mathcal{F}$), whose formal existence is akin to that of the exchange-correlation functional in density functional theory. While the exact dependence of $\mathcal{F}$ on $\mathcal{W}$ is unknown, a few approximate parametric models have been proposed in the past. Here, for a dataset of 923 one-dimensional external potentials with two interacting electrons, we apply machine learning to model $\mathcal{F}$ within the kernel Ansatz. We deal with over-fitting of the kernel to a specific region of phase-space by a one-step regularization not depending on any hyperparameters. Reference correlation energies have been computed by performing exact and Hartree--Fock calculations using discrete variable representation. The resulting models require $\mathcal{W}$ calculated at the Hartree--Fock level as input while yielding monotonous decay in the predicted correlation energies of new molecules reaching sub-chemical accuracy with training.