https://arxiv.org/abs/2405.09622 (or view on SciRate)
The estimation of parameters of a quantum state is a routine component of several quantum protocols across computation, communication and cryptography. In all such estimation tasks, the ultimate limits of precision can be attained by quantum theory by incorporating entanglement in the process. However, multi-copy entangling measurements, which form a key part of this process, are extremely challenging to implement in practice, as reflected in experimental demonstrations to date. We quantify the extent to which precision is lost by performing non-entangling measurements instead, showing this to be limited by the dimension of the quantum system being probed and the number of parameters being estimated.
Entanglement as a resource has been leveraged to enhance quantum effects across computing, communications, sensing, and metrology. And whereas the advantages of entangled states are well established in these domains, the role of entangling measurements, particularly in quantum metrology, remains less understood. Our paper bridges this gap by introducing a quantitative measure of the metrological enhancement provided by entangling measurements and providing simple upper bounds on this key quantity. The Holevo Cramér-Rao bound (HCRB) is the most widely used tool to quantify “attainable” precision in quantum metrology. However, attaining the HCRB requires entangling measurements on infinitely-many copies of the probe state, a resource-intensive demand far beyond the current state-of-the-art, which is limited to two-copy and three-copy entangling measurements. By identifying a situation where such entangling measurements are the most superior to non-entangling ones, we show that the precision prescribed by the HCRB can be arbitrarily far from the practically-attainable precision, where only a few or even none of the copies are entangled before measurement.
We show that for estimating $n$ parameters, the advantage of entangling measurements is at most a factor of $n$ and that for estimating parameters of a $d$-dimensional state, the advantage is at most a factor of $d+1$. These results clarify both the utility of entangling measurements, as well as the limitations of the HCRB, revealing that the HCRB can be arbitrarily far from the practically-attainable precision and questioning the utility of numerous studies focused solely on the HCRB. On the flip side, we prove that were these highly-entangling measurements to be realised, the boost in precision would be arbitrarily large, limited only by system dimension $d$ and number of parameters $n$. Our work advocates a judicious choice of measurement strategy and corresponding performance metric when designing practical quantum estimation tasks or when benchmarking the performance of real-world quantum measurements. By revealing the potential of entangling measurements while also highlighting the practical limitations of the most widely-used attainable-precision bound for them, we connect ideal performance metrics with real-world experimental constraints, bridging the gap between theory and practice. Witnessing the recent surge in experimental demonstrations of multi-copy entangling measurements for quantum estimation, we contribute to the ongoing dialogue within the community by urging researchers to make more informed choices regarding measurement strategies for emerging quantum technologies.