Estimating dissipation from single stationary trajectories

TL;DR

This paper introduces a method to estimate entropy production from stationary time series by analyzing time reversal asymmetry, enabling detection of non-equilibrium processes without detailed mechanistic information.

Contribution

It develops estimators that identify entropy production and non-equilibrium behavior solely from stationary time series data, without requiring flow measurements.

Findings

Estimators successfully detect non-equilibrium processes.

Time reversal asymmetry correlates with entropy production.

Method applicable even with no measurable flows.

Abstract

In this Letter we show that the time reversal asymmetry of a stationary time series provides information about the entropy production of the physical mechanism generating the series, even if one ignores any detail of that mechanism. We develop estimators for the entropy production which can detect non-equilibrium processes even when there are no measurable flows in the time series.