A state with Gaussian characteristic function is a Gaussian state
ProvedBosonicLearn.gaussianAt_of_characteristicThe auxiliary clauses of the definition of a Gaussian state are automatic.
Let be a state on the -mode Fock space whose characteristic function is
for a symmetric real matrix and a vector . Then has finite mean photon number, the series defining its characteristic function converge, and its first and second moments are and : all clauses of the published predicate GaussianAt ρ d γ hold.
Role. IsGaussianState ρ μ Σ is GaussianAt with ; this lemma reduces it to the characteristic function identity, so that states built by displacement or Gaussian mixing are recognized as Gaussian (BosonicLearn.gaussianState_exists_of_le).
Formalization Note GaussianAt and CompleteMoments are the published predicates of Support50.P14 (interleaved quadrature coordinates).
import Mathlib import Definitions.Def_BosonicLearn_husimiQ import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P14 open QAlgorithms.Papers.Independent.Dependency50.P33 (State traceSize) open QAlgorithms.Papers.Independent.Claimed50.P47 (Occupation) open QAlgorithms.Papers.Independent.Support50.P14 open MeasureTheory open scoped ENNReal
namespace BosonicLearn
/-- **A state with a Gaussian characteristic function is a Gaussian state.**
Chen, Anna Mele, Anna Mele, Preskill, *When are bosonic Gaussian states classical to learn?*, arXiv:2609.26705, §1.1, Eqs. (2) and (4), pp. 4–5: "A Gaussian state is a state whose Wigner function is a Gaussian distribution … Such a state
is completely determined by its first and second moments".
Let `ρ` be a state whose characteristic function is `Tr(ρ e^{iξ·r̂}) = exp(−¼ ξᵀγξ + i ξ·d)` for a symmetric real
matrix `γ` and a vector `d`. Then `ρ` satisfies all clauses of the published predicate `GaussianAt ρ d γ`: finite mean
photon number, convergence of the characteristic series, and the first and second moment identities.
**Formalization Note.** `GaussianAt` (`Support50.P14`) is the predicate behind `IsGaussianState`; this lemma says
that its auxiliary clauses (`CompleteMoments`) follow from the characteristic function identity for a state. -/
theorem gaussianAt_of_characteristic {n : ℕ} (ρ : State (Occupation (Fin n)))
(d : Fin n × Fin 2 → ℝ) (γ : Matrix (Fin n × Fin 2) (Fin n × Fin 2) ℝ) (hγ : γ.transpose = γ)
(h : ∀ ξ : Fin n × Fin 2 → ℝ, oscillatorCharacteristic ρ.coefficients ξ =
Complex.exp (((-varianceForm γ ξ / 4 : ℝ) : ℂ) + Complex.I * (meanForm d ξ : ℂ))) :
GaussianAt ρ.coefficients d γ := by sorry
end BosonicLearn