Multitask machine learning of collective variables for enhanced sampling of reactive molecular dynamics

Ability to discover reactions and predict their rates is key to understanding many computational physics, biology and chemistry problems. Enhanced sampling techniques, such as metadynamics and umbrella sampling, use a low-dimensional reaction coordinate / collective variable (CV) space to accelerate sampling for slow reactions. However, their efficiency relies on the choice of CVs, which requires intuition and many trial-and-error tests.

In this work, we propose a multi-task machine learning algorithm to learn collective variables (CVs) from short MD trajectories and transition path sampling, which preserves the information on state labels and potential energies. We show that the algorithm can accurately measure reaction progress and identify the slow motion of the system. It offers a unified dimensionality reduction framework that integrates different learning objectives and can be applied to a wide variety of reactive systems, including Muller-Brown potential model and alanine dipeptide.