Fisher information of correlated stochastic processes

• Published in: New journal of physics 2023-05, Vol.25 (5), p.53037
• Main Authors: Radaelli, Marco, Landi, Gabriel T, Modi, Kavan, Binder, Felix C
• Format: Article
• Language: English
• Subjects: Asymptotic methods; Asymptotic properties; Correlation; Ferromagnetism; Fisher information; information theory; Ising model; metrology; Parameter estimation; Physics; Probability theory; Random variables; spin chains; Stationary processes; Stochastic models; Stochastic processes
• ISSN: 1367-2630; 1367-2630
• DOI: 10.1088/1367-2630/acd321

Many real-world tasks include some kind of parameter estimation, i.e. the determination of a parameter encoded in a probability distribution. Often, such probability distributions arise from stochastic processes. For a stationary stochastic process with temporal correlations, the random variables that constitute it are identically distributed but not independent. This is the case, for instance, for quantum continuous measurements. In this article, we derive the asymptotic Fisher information rate for a stationary process with finite Markov order. We give a precise expression for this rate which is determined by the process’ conditional distribution up to its Markov order. Second, we demonstrate with suitable examples that correlations may both enhance or hamper the metrological precision. Indeed, unlike for entropic information quantities, in general nothing can be said about the sub- or super-additivity of the joint Fisher information in the presence of correlations. To illustrate our results, we apply them to thermometry on an Ising spin chain, considering nearest-neighbour and next-to-nearest neighbour coupling. In this case, the asymptotic Fisher information rate is directly connected to the specific heat capacity of the spin chain. We observe that the presence of correlations strongly enhances the estimation precision in an anti-ferromagnetic chain, while in a ferromagnetic chain this is not the case.