Second-order asymptotics for quantum hypothesis testing in settings beyond i.i.d. - quantum lattice systems and more

TL;DR

This paper extends the second-order asymptotic analysis of quantum hypothesis testing beyond the i.i.d. setting, applying it to quantum lattice systems like Gibbs states at high temperatures, advancing quantum statistical methods.

Contribution

It generalizes second-order asymptotics for quantum hypothesis testing to non-i.i.d. settings such as quantum lattice systems, beyond the traditional i.i.d. framework.

Findings

Extended second-order asymptotics to quantum lattice systems.

Applied results to Gibbs states of quantum spin systems.

Demonstrated convergence rates in non-i.i.d. quantum settings.

Abstract

Quantum Stein's Lemma is a cornerstone of quantum statistics and concerns the problem of correctly identifying a quantum state, given the knowledge that it is one of two specific states ($\rho$ or $\sigma$). It was originally derived in the asymptotic i.i.d. setting, in which arbitrarily many (say, $n$) identical copies of the state ($\rho^{\otimes n}$ or $\sigma^{\otimes n}$) are considered to be available. In this setting, the lemma states that, for any given upper bound on the probability $\alpha_n$ of erroneously inferring the state to be $\sigma$, the probability $\beta_n$ of erroneously inferring the state to be $\rho$ decays exponentially in $n$, with the rate of decay converging to the relative entropy of the two states. The second order asymptotics for quantum hypothesis testing, which establishes the speed of convergence of this rate of decay to its limiting value, was derived in the i.i.d. setting independently by Tomamichel and Hayashi, and Li. We extend this result to settings beyond i.i.d.. Examples of these include Gibbs states of quantum spin systems (with finite-range, translation-invariant interactions) at high temperatures.