Deep-learning approach for the atomic configuration interaction problem on large basis sets

TL;DR

This paper introduces a deep-learning method using convolutional neural networks to efficiently select relevant configurations in large basis sets for atomic structure calculations, significantly reducing computational complexity while maintaining accuracy.

Contribution

The authors develop a convolutional neural network-based approach to preselect configurations in large CI basis sets, enabling accurate atomic calculations that were previously computationally infeasible.

Findings

The method achieves targeted energy precision with fewer configurations.

It outperforms dense neural architectures in accounting for physical basis set structures.

Successfully applied to large basis sets where direct CI is impossible.

Abstract

High-precision atomic structure calculations require accurate modelling of electronic correlations typically addressed via the configuration interaction (CI) problem on a multiconfiguration wave function expansion. The latter can easily become challenging or infeasibly large even for advanced supercomputers. Here we develop a deep-learning approach which allows to preselect the most relevant configurations out of large CI basis sets until the targeted energy precision is achieved. The large CI computation is thereby replaced by a series of smaller ones performed on an iteratively expanding basis subset managed by a neural network. While dense architectures as used in quantum chemistry fail, we show that a convolutional neural network naturally accounts for the physical structure of the basis set and allows for robust and accurate CI calculations. The method was benchmarked on basis sets of moderate size allowing for the direct CI calculation, and further demonstrated on prohibitively large sets where the direct computation is not possible.