Nonequilibrium Enhanced Classical Measurement and Estimation

Zeng Qian,Wang Jin

DOI: https://doi.org/10.1007/s10955-022-02974-8

2022-01-01

Journal of Statistical Physics

Abstract:We consider the classical measurement model for exploring nonequilibrium effects on the physical measurement and information estimation. With appropriate settings on the stochastic dynamics of the measurement and information estimation, we demonstrate the intrinsic statistical correspondence between the entropy production in the measurement and the best estimation accuracy-the Fisher information of the probability distribution of the observed object states. We uncover the general relationship between the moment generating functions of the entropy production and the acquired information from the measurement, which contains the fluctuation information of full orders compared to the Landauer principle on average level. With these relationships, we prove the limit of the best estimation accuracy, which is given by the entropy production in the measurement. This gives rise to a nonequilibrium constraint on the noise-signal-ratio of the estimation, which can be shown as an inequality that is analogous to the thermodynamics uncertainty relation. In applications, when the fluctuation in the observed data is large, the inequality relationships between the moment generating functions lead to a more suitable estimation of the minimum entropy production in the measurement, compared to the Landauer principle; and give rise to a more robust prediction of the best estimation accuracy, compared to the Fisher information. We demonstrate that the entropy production, acquired information and Fisher information are convex functions of the nonequilibrium strength of the measurement dynamics. Thus, the entropy production can be enlarged as the nonequilibrium strength increases in the measurement. As more entropy production is paid, more useful information can be extracted from the measurement, then more accuracy can be achieved in the related estimation.