Markov state modeling is a powerful and widely used technique to reduce complexity and gain insight in large thermal equilibrium molecular simulations. We have developed an extension of this technique to the study of the out of equilibrium dynamics of complex systems . Starting with minimal knowledge of the system dynamics, our approach is effective in reducing high dimensional time evolutions to a Markov process over a handful of relevant distributions in configuration space. These distributions, representing the slow dynamical modes of the system, can be directly used to highlight the relevance of the original degrees of freedom throughout the evolution.
I will highlight the key aspects of this method through its applications to simple and more complex nanotribological models and provide some insight in its future developments towards an automated feature extraction tool for dynamical simulations.
 FP, F. Landes, A. Laio, S. Prestipino, and E. Tosatti, Phys. Rev. E 94, 053001 (2016).