# Mapping Still Matters: Coarse-Graining with Machine Learning Potentials

## TL;DR

This paper explores how different ways of grouping atoms in simulations affect the accuracy of machine learning models used in coarse-grained molecular modeling.

## Contribution

The study introduces practical guidelines for selecting effective coarse-grained mappings compatible with modern machine learning potentials.

## Key findings

- Unphysical bond permutations occur when bonded and nonbonded interaction length scales overlap.
- Encoding species and stereochemistry is crucial to avoid unphysical symmetries.
- The work provides guidance for building transferable coarse-grained models using machine learning.

## Abstract

Coarse-grained (CG) modeling enables molecular simulations
to reach
time and length scales inaccessible to fully atomistic methods. For
classical CG models, the choice of mapping, that is, how atoms are
grouped into CG sites, is a major determinant of accuracy and transferability.
At the same time, the emergence of machine learning potentials (MLPs)
offers new opportunities to build CG models that can in principle
learn the true potential of the mean force for any mapping. In this
work, we systematically investigate how the choice of mapping influences
the representations learned by equivariant MLPs by studying liquid
hexane, amino acids, and polyalanine. We find that when the length
scales of bonded and nonbonded interactions overlap, unphysical bond
permutations can occur. We also demonstrate that correctly encoding
species and maintaining stereochemistry are crucial, as neglecting
either introduces unphysical symmetries. Our findings provide practical
guidance for selecting CG mappings compatible with modern architectures
and guide the development of transferable CG models.