Density functionals from deep learning

TL;DR

This paper introduces a deep learning-based model to approximate the density functional in density-functional theory, enabling more flexible and powerful predictions of quantum system properties.

Contribution

It develops a novel deep learning approach to approximate the density functional, surpassing traditional methods in flexibility and accuracy.

Findings

Performs well on kinetic-energy density functional approximation

Analyzes the model's structure and advantages over traditional machine learning

Demonstrates potential for improved quantum system modeling

Abstract

Density-functional theory is a formally exact description of a many-body quantum system in terms of its density; in practice, however, approximations to the universal density functional are required. In this work, a model based on deep learning is developed to approximate this functional. Deep learning allows computational models that are capable of naturally discovering intricate structure in large and/or high-dimensional data sets, with multiple levels of abstraction. As no assumptions are made as to the form of this structure, this approach is much more powerful and flexible than traditional approaches. As an example application, the model is shown to perform well on approximating the kinetic-energy density functional for noninteracting electrons. The model is analyzed in detail, and its advantages over conventional machine learning are discussed.