Channel Estimation Theory of Low-Noise Multiple Parameters:Attainablity Problem of the Cram{\'e}r-Rao Bounds
Masahiro Hotta, Tokishiro Karasawa

TL;DR
This paper investigates the attainability of the quantum Cramér-Rao bound in low-noise quantum channels with multiple parameters, showing conditions under which the bound can be achieved in the first order, especially with entanglement and non-local measurements.
Contribution
It formulates the multi-parameter low-noise quantum channel and establishes conditions for the simultaneous attainability of the Cramér-Rao bound, including the role of entanglement and system dimension.
Findings
Cramér-Rao bound can be attained for low-noise channels when D ≤ N-1.
Entanglement with ancillas extends attainability to D ≤ N^2 - 1.
First-order attainability is generally possible despite non-commutativity issues.
Abstract
For decoherence processes induced by weak interactions with the environment, a general quantum channel with one noise parameter has been formulated. This channel is called low-noise channel and very useful for investigating the parameter estimation in the leading order. In this paper, we formulate the low-noise channel with multiple unknown parameters in order to address the simultaneous achievability of the Cram{\'e}r-Rao bound for the parameters estimation. In general, the simultaneous achievement of the Cram{\'e}r-Rao bound for multi-parameter estimations suffers from non-commutativity of optimal measurements for respective parameters. However, with certain exceptions, we show that the Cram{\'e}r-Rao bound for output states of dissipative low-noise channels can be always attained in the first order of the parameters as long as D \leq N-1, where D and N denote the number of the…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
