Thermodynamic Interpolation: A generative approach to molecular thermodynamics and kinetics

TL;DR

This paper introduces Thermodynamic Interpolation (TI), a novel method using normalizing flows for temperature-controlled sampling in molecular thermodynamics, enabling extrapolation and improved free energy and kinetic rate estimations.

Contribution

The work presents a new ambient-space approach for thermodynamic interpolation that works directly in configurational space, allowing flexible temperature control and extrapolation beyond training data.

Findings

TI effectively generates sampling statistics across temperatures.

The method accurately estimates free energy differences.

It enables kinetic rate estimation via generator extended dynamic mode decomposition.

Abstract

Using normalizing flows and reweighting, Boltzmann Generators enable equilibrium sampling from a Boltzmann distribution, defined by an energy function and thermodynamic state. In this work, we introduce Thermodynamic Interpolation (TI), which allows for generating sampling statistics in a temperature-controllable way. We introduce TI flavors that work directly in the ambient configurational space, mapping between different thermodynamic states or through a latent, normally distributed reference state. Our ambient-space approach allows for the specification of arbitrary target temperatures, ensuring generalizability within the temperature range of the training set and demonstrating the potential for extrapolation beyond it. We validate the effectiveness of TI on model systems that exhibit metastability and non-trivial temperature dependencies. Finally, we demonstrate how to combine TI-based sampling to estimate free energy differences through various free energy perturbation methods and provide corresponding approximated kinetic rates estimated through generator extended dynamic mode decomposition (gEDMD).