Proud to share my first paper in collaboration with J. Borregaard and @jenseisert. We provide a variational quantum algorithm for the optimization of noisy quantum multi-parameter estimation. scirate.com/arxiv/2006.063…
This work provides a new application of quantum technologies (both computers and other platforms) to optimize the next generation of quantum metrology devices. This mindset was recently coined as "quantum-computer aided design" by @A_Aspuru_Guzik et al.
See below for a summary of the paper 👇 
In quantum metrology, a probe state is prepared and then undergoes a noisy evolution encoding parameters (e.g. phases). The resulting state is then measured to obtain a classical probability distribution from which the parameters or a function thereof is estimated.
We parametrize both the probe state and the measurement and compute a cost function from the Cramér-Rao bound that quantifies the estimation precision. We use a classical feedback loop to reduce the cost and thus improve the possible estimation precision.
Using @pennylaneai, we apply our algorithm to the estimation of phase averages in a Ramsay interferometry-like setting under dephasing. We reproduce the known advantage of GHZ states in the noisefree case and find protocols improving over standard methods in the noisy case.
As a new application, we consider spin location sensing using NV centers under dephasing. We study the trade-off between a local and a specific shallow entangled probe preparation when gates are depolarizing. The entanglement-advantage disappears as soon as the noise is non-zero.
We use the parameter-shift rule to compute necessary derivatives and observe that it also holds in the noisy case and provide additional shift rules for convex combinations of quantum channels. Appendix A can be read as a standalone intro to this.
For a demo on noisy gradients, see the excellent demo from the @pennylaneai repository: pennylane.ai/qml/demos/tuto…
In Appendix D, we provide extensions of the method to include prior knowledge, account for mutual time-dependence of parameter and noise and show how to benchmark against the Quantum Cramér-Rao bound in the noisefree case.
We endorse scientific-conduct.github.io and provide a CO2 emission table to raise awareness for the climate impact of research.
TL;DR: Quantum computers can now be used to improve quantum metrology!
#quantumtwitter #quantum #quantumtech
#quantumtwitter #quantum #quantumtech
@thereaderapp unroll
@threadreaderapp unroll
