Multiscale reweighted stochastic embedding (MRSE): Deep learning of collective variables for enhanced sampling

TL;DR

This paper introduces MRSE, a deep learning method that automatically learns effective collective variables for enhanced sampling in molecular simulations, improving the identification of metastable states.

Contribution

The paper presents MRSE, a novel multiscale reweighted stochastic embedding technique that advances CV learning by integrating weighted sampling, multiscale modeling, and reweighting for biased data.

Findings

Successfully characterizes metastable states in model systems

Constructs low-dimensional CVs that distinguish different states

Enhances sampling efficiency in molecular simulations

Abstract

Machine learning methods provide a general framework for automatically finding and representing the essential characteristics of simulation data. This task is particularly crucial in enhanced sampling simulations. There we seek a few generalized degrees of freedom, referred to as collective variables (CVs), to represent and drive the sampling of the free energy landscape. In theory, these CVs should separate different metastable states and correspond to the slow degrees of freedom of the studied physical process. To this aim, we propose a new method that we call multiscale reweighted stochastic embedding (MRSE). Our work builds upon a parametric version of stochastic neighbor embedding. The technique automatically learns CVs that map a high-dimensional feature space to a low-dimensional latent space via a deep neural network. We introduce several new advancements to stochastic neighbor embedding methods that make MRSE especially suitable for enhanced sampling simulations: (1) weight-tempered random sampling as a landmark selection scheme to obtain training data sets that strike a balance between equilibrium representation and capturing important metastable states lying higher in free energy; (2) a multiscale representation of the high-dimensional feature space via a Gaussian mixture probability model; and (3) a reweighting procedure to account for training data from a biased probability distribution. We show that MRSE constructs low-dimensional CVs that can correctly characterize the different metastable states in three model systems: the M\"uller-Brown potential, alanine dipeptide, and alanine tetrapeptide.