Quantum algorithms for matrix geometric means
Nana Liu,Qisheng Wang,Mark M. Wilde,Zhicheng Zhang
2024-05-02
Abstract:Matrix geometric means between two positive definite matrices can be defined equivalently from distinct perspectives - as solutions to certain nonlinear systems of equations, as points along geodesics in Riemannian geometry, and as solutions to certain optimisation problems. This diversity already suggests the potential for varied applications, as well as acting as a bridge between different domains. Here we devise new quantum subroutines to efficiently prepare quantum unitary operators that embed the standard matrix geometric mean and its generalisations called the weighted matrix geometric mean. This enables the construction of solutions to the algebraic Riccati equation, which is an important class of nonlinear systems of equations that appears in machine learning, optimal control, estimation, and filtering. Using these subroutines, we present a new class of quantum learning algorithms called quantum geometric mean metric learning. This has applications in efficiently finding the best distance measure and solving classification problems in the weakly supervised limit and for anomaly detection, for both classical and quantum problems. We also show how our method can be generalised to a particular p^th-order system of nonlinear equations. These quantum subroutines for matrix geometric means are also useful in other areas of quantum information. For example, we show how to use them in the estimation of geometric Renyi relative entropies and the Uhlmann fidelity by means of the Fuchs-Caves observable. In particular, our quantum algorithms for estimating the Uhlmann and Matsumoto fidelities have optimal dependence on the precision. Finally, we provide a BQP-complete problem based on matrix geometric means that can be solved by our subroutines, thus characterising their computational capability.
Quantum Physics,Mathematical Physics,Functional Analysis,Operator Algebras