# Techniques: Building trust for continuous variable quantum states

## What the paper contributes (one paragraph)
The paper introduces a single receive-and-measure protocol for characterising continuous variable quantum states using only heterodyne (double-homodyne) detection, a Gaussian measurement. Its technical core is an analytical estimator for the expected value of any operator with bounded support over the Fock basis, derived from an approximate optical equivalence theorem and computed directly from heterodyne samples, with an explicit energy-dependent error bound. The same measurement is reused for three tasks that differ only in the classical post-processing and the number of subsystems sampled. Under an i.i.d. assumption it gives reliable heterodyne tomography with analytical confidence intervals on density-matrix elements, needing no binning or data reconstruction. Dropping bounded support but keeping i.i.d. gives a certification protocol that estimates fidelity to a pure target. Dropping i.i.d. entirely, via an infinite-dimensional de Finetti reduction against permutation-invariant states, gives a verification protocol against a fully malicious prover.

## Techniques offered
- **Reliable heterodyne state tomography** - reconstructs the density-matrix elements of an i.i.d. continuous variable state directly from heterodyne samples using a bounded estimator, with no binning or mathematical reconstruction and a single fixed Gaussian measurement setting.
  - guarantee: statistical bound; analytical confidence intervals on each density-matrix element with sample count scaling as poly(1/epsilon, log(1/delta)).
  - quote: "This protocol allows to perform reliable continuous variable quantum state tomography based on heterodyne detection, with analytical confidence intervals, which we refer to as \textit{heterodyne tomography} in what follows."
- **Gaussian state certification** - estimates the fidelity between an i.i.d. (no longer bounded-support) source state and a chosen pure target from heterodyne samples plus a support-estimation step.
  - guarantee: statistical bound under i.i.d. assumption; bounds the joint probability that the fidelity estimate is inaccurate and the support test passes, with sample count poly(m, 1/epsilon, 1/delta).
  - quote: "We then derive a \textit{receive-and-measure} certification protocol (against i.i.d.\@ adversary) for continuous variable quantum states, with Gaussian measurements (Theorem~\ref{thi.i.d.})."
- **Gaussian verification against a malicious prover** - promotes the certification protocol to verify many copies of a pure target state with no i.i.d. assumption, using an infinite-dimensional de Finetti reduction to restrict the adversary to almost-i.i.d. behaviour.
  - guarantee: interactive soundness; with high probability either the fidelity estimate is accurate or the support test fails, against an arbitrary prover, with parameters poly(m).
  - quote: "We further promote this certification technique to a verification protocol against fully malicious adversary  (Theorem~\ref{thVUCVQC}), using a de Finetti reduction for infinite-dimensional systems~\cite{renner2009finetti}."

## Where it could apply
- Target primitive(s): sampling (heterodyne / Gaussian boson sampling output states), continuous variable state preparation, delegated continuous variable computation; any pure target state with bounded support over the Fock basis
- Target application group(s): platforms_tooling, error_correction (verification and certification of device output), crypto_security (delegated computation, adversarial verification)

## Caveats
The fidelity estimator targets pure target states only (mixed targets are left as future work, and the operator estimator does not yet extend to quantum maps). Efficiency is polynomial in the number of copies m but the constants and energy parameters (E, s) must be tuned per target state, and the bounded-support assumption only holds in the trusted tomography setting. The work is theoretical with no experimental implementation; no paywall or access limits.