DeepLNE++ leveraging knowledge distillation for accelerated multi-state path-like collective variables

TL;DR

DeepLNE++ is an improved machine learning method that accelerates the computation of path-like collective variables in molecular dynamics, enabling efficient modeling of complex biomolecular processes.

Contribution

It introduces a knowledge distillation approach to speed up DeepLNE and incorporates system-specific knowledge for better performance.

Findings

Significantly faster evaluation of collective variables.

Enables free energy landscape computation for large biomolecular systems.

Enhanced versatility through system-specific knowledge encoding.

Abstract

Path-like collective variables can be very effective for accurately modeling complex biomolecular processes in molecular dynamics simulations. Recently, we introduced DeepLNE, a machine learning-based path-like CV that provides a progression variable s along the path as a non-linear combination of several descriptors, effectively approximating the reaction coordinate. However, DeepLNE is computationally expensive for realistic systems needing many descriptors and limited in its ability to handle multi-state reactions. Here we present DeepLNE++, which uses a knowledge distillation approach to significantly accelerate the evaluation of DeepLNE, making it feasible to compute free energy landscapes for large and complex biomolecular systems. In addition, DeepLNE++ encodes system-specific knowledge within a supervised multitasking framework, enhancing its versatility and effectiveness.