Descriptors-free Collective Variables From Geometric Graph Neural Networks

TL;DR

This paper introduces a fully automatic, symmetry-invariant method for defining collective variables in enhanced sampling simulations using geometric graph neural networks, eliminating the need for manual feature selection.

Contribution

It presents a novel graph neural network-based approach that directly uses atomic coordinates to determine collective variables without predefined descriptors.

Findings

Method is robust across different systems.

Achieves automatic CV determination invariant under symmetries.

Proven effective on peptides, ion dissociation, and chemical reactions.

Abstract

Enhanced sampling simulations make the computational study of rare events feasible. A large family of such methods crucially depends on the definition of some collective variables (CVs) that could provide a low-dimensional representation of the relevant physics of the process. Recently, many methods have been proposed to semi-automatize the CV design by using machine learning tools to learn the variables directly from the simulation data. However, most methods are based on feed-forward neural networks and require as input some user-defined physical descriptors. Here, we propose to bypass this step using a graph neural network to directly use the atomic coordinates as input for the CV model. This way, we achieve a fully automatic approach to CV determination that provides variables invariant under the relevant symmetries, especially the permutational one. Furthermore, we provide different analysis tools to favor the physical interpretation of the final CV. We prove the robustness of our approach using different methods from the literature for the optimization of the CV, and we prove its efficacy on several systems, including a small peptide, an ion dissociation in explicit solvent, and a simple chemical reaction.