Efficient computation of average subsystem Bures distance between fermionic Gaussian states

Abstract

The average subsystem trace distance has been proposed as an indicator of quantum many-body chaos and integrability. However, evaluating it presents two main difficulties: high computational cost for large systems and ambiguities in defining and ordering eigenstates in integrable systems. In this work, we develop an efficient algorithm to compute the Bures distance between fermionic Gaussian states, enabling access to larger system sizes. Using this method, we calculate the average subsystem Bures distance for eigenstates in the spin-1/2 transverse-field Ising chain and the Dirac fermion formulation of the quadratic Sachdev-Ye-Kitaev (Dirac SYK$_2$) model, as well as for random pure fermionic Gaussian states. To handle degeneracy in the Ising chain, we consider simultaneous eigenstates of all local conserved charges and employ these charges to systematically order degenerate states. Our results are consistent with the earlier conjecture of a linear growth with subsystem size. We show that the distinct scaling of the average subsystem distances in chaotic versus integrable systems originates from discontinuities of local conserved charges across the spectrum in integrable models. For the Dirac SYK$_2$ model and random pure Gaussian states, we obtain similar results for the average subsystem distances, which do not show a linear increase.