arXiv:2610.24686
Efficient Quantum Algorithm for Mixed-State Quantum Geometry
Chenyang Wang, Qi Zhao, Xiao Yuan, Xiaoxia Cai
Abstract
Mixed-state quantum geometry plays a fundamental role in quantum metrology, multiparameter estimation, and the characterization of quantum many-body systems. Yet, unlike its pure-state counterpart, accessing mixed-state geometry requires solving an ill-conditioned Lyapunov equation for the symmetric logarithmic derivative (SLD), posing a major challenge for resource-efficient quantum computation. Here we address the challenge and develop a novel quantum algorithmic framework for efficiently accessing the quantum Fisher information matrix (QFIM) and quantum geometric tensor (QGT) of mixed states. Our key insight is to uncover a previously unexplored connection between the positive semidefinite structure of density matrices and exponentially convergent Gauss-Laguerre quadrature, enabling an efficient integral-based construction of the SLD. A rigorous complexity analysis further establishes the resource efficiency of our algorithm, demonstrating logarithmic precision dependence and double-logarithmic ancillary-qubit overhead at fixed conditioning. Building on this construction, we achieve near-Heisenberg precision scaling for diagonal QFI estimation without quantum phase estimation and construct unified block encodings of the full QFIM and QGT under a bounded spectral-gap assumption. A complementary regularized algorithm extends efficient diagonal estimation to exponentially small spectral gaps under a bounded-response condition. Our results establish a resource-efficient route to coherent mixed-state quantum geometry, enabling its direct integration into quantum linear-algebra algorithms without classical reconstruction and opening new possibilities for quantum-enhanced multiparameter metrology.
AI Summary
The paper addresses quantum access to mixed-state quantum geometry, particularly the quantum Fisher information matrix (QFIM) and quantum geometric tensor (QGT). Computing these quantities involves the symmetric logarithmic derivative (SLD), defined through a Lyapunov equation that can become ill-conditioned when the density matrix has small eigenvalues. The proposed framework constructs the SLD using an integral representation and Gauss–Laguerre quadrature. It connects the positive-semidefinite structure of density matrices to exponential quadrature convergence. The abstract reports a rigorous complexity analysis giving logarithmic dependence on inverse precision and double-logarithmic ancillary-qubit overhead at fixed conditioning. Building on this construction, the framework estimates diagonal quantum Fisher information with near-Heisenberg precision scaling, without quantum phase estimation. It also provides unified block encodings of the full QFIM and QGT, intended to make these matrices available coherently to quantum linear-algebra algorithms rather than requiring classical reconstruction. The stated guarantees have distinct conditions: full-matrix block encodings assume a bounded spectral gap, while a complementary regularized algorithm permits efficient diagonal estimation with exponentially small spectral gaps under a bounded-response condition. The abstract does not specify the precise gap or response definitions, access-model assumptions, or complete complexity bounds. Consequently, the reported precision and ancilla scalings cannot be assessed independently of conditioning and input-access costs from the available text.
Written by gpt-6.1-sol from the abstract only. Automated: it can be wrong.