Scalable learning of potentials to predict time-dependent Hartree-Fock dynamics

TL;DR

This paper introduces a scalable machine learning framework to accurately predict the inter-electronic potential in time-dependent Hartree-Fock simulations, facilitating improved quantum dynamics modeling for molecules.

Contribution

It develops and tests three models for learning the TDHF inter-electronic potential, incorporating symmetry considerations to enhance performance and scalability for larger molecular systems.

Findings

Models with eight-fold symmetry outperform others in accuracy and efficiency.

All models accurately predict electron dynamics even outside training conditions.

The approach enables matrix-free training suitable for large molecules.

Abstract

We propose a framework to learn the time-dependent Hartree-Fock (TDHF) inter-electronic potential of a molecule from its electron density dynamics. Though the entire TDHF Hamiltonian, including the inter-electronic potential, can be computed from first principles, we use this problem as a testbed to develop strategies that can be applied to learn a priori unknown terms that arise in other methods/approaches to quantum dynamics, e.g., emerging problems such as learning exchange-correlation potentials for time-dependent density functional theory. We develop, train, and test three models of the TDHF inter-electronic potential, each parameterized by a four-index tensor of size up to $60 \times 60 \times 60 \times 60$. Two of the models preserve Hermitian symmetry, while one model preserves an eight-fold permutation symmetry that implies Hermitian symmetry. Across seven different molecular systems, we find that accounting for the deeper eight-fold symmetry leads to the best-performing model across three metrics: training efficiency, test set predictive power, and direct comparison of true and learned inter-electronic potentials. All three models, when trained on ensembles of field-free trajectories, generate accurate electron dynamics predictions even in a field-on regime that lies outside the training set. To enable our models to scale to large molecular systems, we derive expressions for Jacobian-vector products that enable iterative, matrix-free training.