Unnumbered local Gaussification construction, (5)-(11)
OpenQAlgorithms.Papers.Independent.Support50.P14.localGaussificationLet be a normalized density on a nonempty finite collection of full oscillator modes, with zero first moments and finite symmetric covariance . For , label the copies by binary strings and use the Sylvester matrix . There exist bounded unitaries implementing on both quadratures of the copies at each site. Their action on gives normalized states with computed one-copy marginals . A normalized Gaussian density with moments exists, and
Each marginal has the same moments and characteristic function , where for the zero row and otherwise. This construction supplies the local Gaussian limit used by Lemma 1.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P14
noncomputable section
open Filter
open scoped BigOperators ENNReal Topology
attribute [local instance] Classical.propDecidable Classical.decEq
namespace QAlgorithms.Papers.Independent.Support50.P14
open QAlgorithms.RealRegisters (Hilbert)
open QAlgorithms.Papers.Independent.Dependency50.P33
(State IsDensity IsTraceClass traceSize)
open QAlgorithms.Papers.Independent.Dependency50.P39 (displacementMatrix)
open QAlgorithms.Papers.Independent.Frontier30.P14
(unitaryCoeff conjugateKernel boundedUnitaryConjugation)
open QAlgorithms.Papers.Independent.Frontier18.P07 (extendedEntropyNats)
open QAlgorithms.Papers.Independent.Claimed50.P47 (Occupation)
open QAlgorithms.Papers.Independent.Gap48.P20 (Copies)
/-- Local Gaussification, PDF p.2, (5)--(11). For a centered finite-energy
input, actual bounded Sylvester copy mixers and their computed normalized
marginals converge in trace norm UNIFORMLY in the growing copy label.
The exact two characteristic-function formulas are retained. -/
theorem localGaussification {I : Type} [Fintype I] [Nonempty I]
(rho : State (Occupation I)) (gamma : OscillatorCovariance I)
(hmoments : CompleteMoments rho.coefficients 0 gamma) :
∃ rhoG : State (Occupation I), GaussianAt rhoG.coefficients 0 gamma ∧
∃ U : (m : ℕ) → Hilbert (Occupation (Copies Bool m)) ≃ₗᵢ[ℂ]
Hilbert (Occupation (Copies Bool m)),
(∀ m (xi : InterleavedQuadrature (Copies Bool m) → ℝ),
boundedUnitaryConjugation (U m).symm (oscillatorWeylKernel xi) = oscillatorWeylKernel (mixedArgument m xi)) ∧
(∀ m, IsDensity (conjugateKernel (copyMixerKernel m (U m))
(copiesKernel m rho.coefficients))) ∧
(∀ m (k : Copies Bool m),
let reduced := copyMarginal (conjugateKernel (copyMixerKernel m (U m))
(copiesKernel m rho.coefficients)) k
IsDensity reduced ∧ CompleteMoments reduced 0 gamma ∧
∀ xi : InterleavedQuadrature I → ℝ,
oscillatorCharacteristic reduced xi =
oscillatorCharacteristic rho.coefficients (((Real.sqrt ((2 : ℝ) ^ m))⁻¹) • xi) ^
positiveEntries m k *
oscillatorCharacteristic rho.coefficients ((-(Real.sqrt ((2 : ℝ) ^ m))⁻¹) • xi) ^
(2 ^ m - positiveEntries m k)) ∧
(∀ epsilon : ℝ, 0 < epsilon → ∃ m0 : ℕ, ∀ m : ℕ, m0 ≤ m →
∀ k : Copies Bool m,
traceSize (copyMarginal (conjugateKernel (copyMixerKernel m (U m))
(copiesKernel m rho.coefficients)) k - rhoG.coefficients) <
ENNReal.ofReal epsilon) := by sorry
end QAlgorithms.Papers.Independent.Support50.P14
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be a finite nonempty type, let be a density matrix indexed by occupations , and let be a real matrix indexed by . Here a density on an index type means a complex matrix admitting an expansion , with and summable costs , such that every finite quadratic form is real and nonnegative, and such that its real diagonal is summable with sum . Write for the infimum in of those expansion costs, taking an empty infimum to be . For any finite mode type , define
Assume that has the following complete moments with zero mean and matrix : , , and for every , each inner family and the outer family of sums defining are summable; furthermore has derivative at , and its total derivative function has derivative there. Then there exists a density on having these same complete moments and satisfying for every . There also exists, for every natural number , a complex linear isometric equivalence of with itself, where , with all the following properties. For every real , define
Writing , where is the coordinate unit vector, the inverse equivalence satisfies the coefficient identity
For occupations , define
and
For every , is a density. For every , define its selected-copy matrix by
where when , and otherwise. Then is a density and has the same complete zero-mean moments and matrix described above, and for every real ,
where the division in is natural-number division. Finally, the convergence is uniform over the selected copy:
All natural numbers , including , are covered: at , there is one copy, , and the second characteristic-function factor has exponent , so equals even when its base is . The coefficient sums used in the definitions are totalized sums, equal to for nonsummable complex families; summability is explicitly required for the characteristic functions in the complete-moment conditions, but is not separately asserted for every conjugation or marginal sum.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.