Integrated VAC: A robust strategy for identifying eigenfunctions of dynamical operators

TL;DR

This paper introduces integrated VAC (IVAC), a robust extension of the variational approach to conformational dynamics that improves eigenfunction estimation by integrating over multiple lag times, addressing VAC's sensitivity to lag time choice and overfitting issues.

Contribution

The paper proposes IVAC, an extension of VAC that enhances robustness and reproducibility by integrating over multiple lag times, mitigating parameter sensitivity and overfitting.

Findings

IVAC provides more stable eigenfunction estimates across different lag times.

IVAC reduces overfitting compared to traditional VAC with neural networks.

IVAC demonstrates improved reproducibility in identifying dynamical patterns.

Abstract

One approach to analyzing the dynamics of a physical system is to search for long-lived patterns in its motions. This approach has been particularly successful for molecular dynamics data, where slowly decorrelating patterns can indicate large-scale conformational changes. Detecting such patterns is the central objective of the variational approach to conformational dynamics (VAC), as well as the related methods of time-lagged independent component analysis and Markov state modeling. In VAC, the search for slowly decorrelating patterns is formalized as a variational problem solved by the eigenfunctions of the system's transition operator. VAC computes solutions to this variational problem by optimizing a linear or nonlinear model of the eigenfunctions using time series data. Here, we build on VAC's success by addressing two practical limitations. First, VAC can give poor eigenfunction estimates when the lag time parameter is chosen poorly. Second, VAC can overfit when using flexible parameterizations such as artificial neural networks with insufficient regularization. To address these issues, we propose an extension that we call integrated VAC (IVAC). IVAC integrates over multiple lag times before solving the variational problem, making its results more robust and reproducible than VAC's.