Strong and uniform convergence in the teleportation simulation of bosonic Gaussian channels

TL;DR

This paper clarifies the convergence properties of bosonic Gaussian channel teleportation simulations, showing strong and uniform convergence under certain conditions and providing explicit bounds, which impact quantum information theory proofs.

Contribution

The paper explicitly distinguishes between strong and uniform convergence in bosonic Gaussian channel teleportation simulations and provides explicit bounds for their accuracy.

Findings

Teleportation convergence is strong but not uniform for general input states.

Pure-loss, thermal, and amplifier channels converge both strongly and uniformly.

Explicit bounds on the accuracy of channel simulations are provided.

Abstract

In the literature on the continuous-variable bosonic teleportation protocol due to [Braunstein and Kimble, Phys. Rev. Lett., 80(4):869, 1998], it is often loosely stated that this protocol converges to a perfect teleportation of an input state in the limit of ideal squeezing and ideal detection, but the exact form of this convergence is typically not clarified. In this paper, I explicitly clarify that the convergence is in the strong sense, and not the uniform sense, and furthermore, that the convergence occurs for any input state to the protocol, including the infinite-energy Basel states defined and discussed here. I also prove, in contrast to the above result, that the teleportation simulations of pure-loss, thermal, pure-amplifier, amplifier, and additive-noise channels converge both strongly and uniformly to the original channels, in the limit of ideal squeezing and detection for the simulations. For these channels, I give explicit uniform bounds on the accuracy of their teleportation simulations. I then extend these uniform convergence results to particular multi-mode bosonic Gaussian channels. These convergence statements have important implications for mathematical proofs that make use of the teleportation simulation of bosonic Gaussian channels, some of which have to do with bounding their non-asymptotic secret-key-agreement capacities. As a byproduct of the discussion given here, I confirm the correctness of the proof of such bounds from my joint work with Berta and Tomamichel from [Wilde, Tomamichel, Berta, IEEE Trans. Inf. Theory 63(3):1792, March 2017]. Furthermore, I show that it is not necessary to invoke the energy-constrained diamond distance in order to confirm the correctness of this proof.