Fast Quantum Algorithms for Trace Distance Estimation

TL;DR

This paper introduces efficient quantum algorithms for estimating the trace distance between mixed quantum states of low rank, achieving complexities independent of state dimension and applicable to quantum state certification.

Contribution

The paper presents novel quantum algorithms for trace distance estimation with complexities independent of quantum state dimension, including a BQP-complete decision problem.

Findings

Algorithms use $r ilde{O}(1/\varepsilon^2)$ queries and $\tilde{O}(r^2/\varepsilon^5)$ samples.

Query and sample complexities are independent of the dimension $N$ of quantum states.

Decision version of low-rank trace distance estimation is BQP-complete.

Abstract

In quantum information, trace distance is a basic metric of distinguishability between quantum states. However, there is no known efficient approach to estimate the value of trace distance in general. In this paper, we propose efficient quantum algorithms for estimating the trace distance within additive error $\varepsilon$ between mixed quantum states of rank $r$. Specifically, we first provide a quantum algorithm using $r \cdot \widetilde O(1/\varepsilon^2)$ queries to the quantum circuits that prepare the purifications of quantum states. Then, we modify this quantum algorithm to obtain another algorithm using $\widetilde O(r^2/\varepsilon^5)$ samples of quantum states, which can be applied to quantum state certification. These algorithms have query/sample complexities that are independent of the dimension $N$ of quantum states, and their time complexities only incur an extra $O(\log (N))$ factor. In addition, we show that the decision version of low-rank trace distance estimation is $\mathsf{BQP}$-complete.