Mutual Information as a Two-Point Correlation Function in Stochastic Lattice Models

In statistical physics entropy is usually introduced as a global quantity which expresses the amount of information that would be needed to specify the microscopic configuration of a system. However, for lattice models with infinitely many possible configurations per lattice site it is also meaningful to introduce entropy as a local observable that describes the information content of a single lattice site. Likewise, the mutual information can be interpreted as a two-point correlation function. Studying a particular growth model we demonstrate that the mutual information exhibits scaling properties that are consistent with the established phenomenological scaling picture.