A measurable family of Gaussian states for classical covariances
ProvedBosonicLearn.gaussianState_classical_familyThe Gaussian states with covariance above the vacuum form a measurable family.
There is a map from pairs (mean , covariance ) to states on the -mode Fock space such that
- is measurable for the Borel structure of the trace-norm topology on states;
- every value is a Gaussian state;
- if is positive semidefinite, then is the Gaussian state .
For the state is the classical mixture of coherent states ; outside this closed set is constant (the vacuum).
Role. The heterodyne tomography protocols of Theorem 8 and of [BMEM25, Theorem 10] output "a valid Gaussian state" computed measurably from the data, with a covariance that dominates the true one; under the promise this lemma provides that output. It replaces BosonicLearn.gaussianState_measurable_family (which asks for all Gaussian states, including squeezed ones) in the decompositions of BosonicLearn.theorem_8 and BosonicLearn.bmem25_theorem_10.
Formalization Note The covariance argument is a function Fin n ⊕ Fin n → Fin n ⊕ Fin n → ℝ, so that the domain carries the product measurable structure; State carries the published trace-norm topology and its Borel σ-algebra.
import Mathlib import Definitions.Def_BosonicLearn_IsGaussianState open QAlgorithms.Papers.Independent.Dependency50.P33 (State traceSize) open QAlgorithms.Papers.Independent.Claimed50.P47 (Occupation) open MeasureTheory open scoped ENNReal Matrix
namespace BosonicLearn
/-- **A measurable family of Gaussian states with classical covariance (`Σ ⪰ ½I`).**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, Theorem 8, p. 61: the protocol "always outputs a valid Gaussian state ρ̂" computed from the heterodyne data; proof, p. 62: the
output is the Gaussian state with the estimated mean and an inflated covariance estimate, which dominates the true
covariance `Σ ⪰ (½ + ν)I ⪰ ½I`.
There is a map `gs` from pairs (mean, covariance) to states on the `n`-mode Fock space that is measurable for the
trace-norm Borel structure, takes only Gaussian states as values, and such that `gs(µ, Σ)` is the Gaussian state
`ρ(µ, Σ)` whenever `Σ − ½I` is positive semidefinite.
**Formalization Note.** For `Σ ⪰ ½I` the Gaussian state is the classical mixture of coherent states
`∫ |α(x)⟩⟨α(x)| dN(µ, Σ − ½I)(x)`, which exists and depends continuously on `(µ, Σ)`; outside this closed set `gs` may
be constant. This is the form of `gaussianState_measurable_family` needed for heterodyne tomography under the
promise `Σ ⪰ ½I`; it avoids squeezed states. The covariance argument is a function `Fin n ⊕ Fin n → Fin n ⊕ Fin n → ℝ`
so that the domain carries the product measurable structure. -/
theorem gaussianState_classical_family (n : ℕ) :
∃ gs : (Fin n ⊕ Fin n → ℝ) × (Fin n ⊕ Fin n → Fin n ⊕ Fin n → ℝ) → State (Occupation (Fin n)),
Measurable gs ∧
(∀ p, ∃ (μ : Fin n ⊕ Fin n → ℝ) (S : Matrix (Fin n ⊕ Fin n) (Fin n ⊕ Fin n) ℝ),
IsGaussianState (gs p) μ S) ∧
∀ (μ : Fin n ⊕ Fin n → ℝ) (S : Matrix (Fin n ⊕ Fin n) (Fin n ⊕ Fin n) ℝ),
(S - (1 / 2 : ℝ) • (1 : Matrix (Fin n ⊕ Fin n) (Fin n ⊕ Fin n) ℝ)).PosSemidef →
IsGaussianState (gs (μ, fun a b => S a b)) μ S := by sorry
end BosonicLearn