q-pac: A Python package for machine learned charge equilibration models.

Area of Science:

• Computational chemistry
• Materials science
• Machine learning

Background:

• Current machine learning interatomic potentials often use local representations, neglecting long-range electrostatic interactions and non-local charge transfer.
• Accurate modeling of electrostatics is crucial for understanding molecular and material behavior in response to external fields.

Purpose of the Study:

• To introduce the q-pac Python package, which enhances the kQEq method for machine learning-based charge equilibration.
• To provide a flexible framework for developing advanced machine learning charge equilibration models.

Main Methods:

• The study implements algorithmic and methodological improvements to the kQEq method, which uses Kernel Machine Learning (Kernel ML) to predict atomic electronegativities.
• The q-pac package facilitates the rigorous calculation of long-range electrostatic interactions.

Main Results:

• The q-pac package offers an extendable framework for machine learning charge equilibration.
• It enables accurate prediction of electrostatic interactions and energy responses.

Conclusions:

• The q-pac package represents a significant step forward in developing robust machine learning models for charge equilibration.
• This work facilitates more accurate simulations of electrostatic effects in molecules and materials.