Accelerating Electron Dynamics Simulations through Machine Learned Time Propagators

TL;DR

This paper introduces a machine learning approach using neural operators to accelerate real-time TDDFT electron dynamics simulations, achieving higher accuracy and speed for modeling laser-irradiated molecules.

Contribution

The work presents a novel neural operator-based time-propagator for TDDFT, integrating physics constraints to improve simulation efficiency and accuracy.

Findings

Achieves faster electron dynamics simulations compared to traditional methods.

Maintains high accuracy with physics-informed neural operators.

Demonstrates effectiveness on diatomic molecules.

Abstract

Time-dependent density functional theory (TDDFT) is a widely used method to investigate electron dynamics under various external perturbations such as laser fields. In this work, we present a novel approach to accelerate real time TDDFT based electron dynamics simulations using autoregressive neural operators as time-propagators for the electron density. By leveraging physics-informed constraints and high-resolution training data, our model achieves superior accuracy and computational speed compared to traditional numerical solvers. We demonstrate the effectiveness of our model on a class of one-dimensional diatomic molecules. This method has potential in enabling real-time, on-the-fly modeling of laser-irradiated molecules and materials with varying experimental parameters.