Minlos Analyticity — Symmetry and Moments for Gaussian Measures #
This file provides infrastructure for Gaussian measures constructed via Minlos' theorem:
CovarianceForm: Real symmetric bilinear form for covariancenegMap,integral_neg_invariance: Symmetry under sign flip (uses Minlos uniqueness)moment_zero_from_realCF: Zero mean from characteristic functional symmetry
Contents #
This file provides infrastructure for Gaussian measures via Minlos. No axioms declared here.
A real symmetric, positive semidefinite covariance form on real test functions, together with a proof that the associated Gaussian characteristic functional exp(-½Q(f,f)) is positive definite (in the bochner sense).
- Q : TestFunction → TestFunction → ℝ
The real covariance bilinear form.
- cont_diag : Continuous fun (f : TestFunction) => self.Q f f
- gaussian_cf_pd : IsPositiveDefinite fun (f : TestFunction) => Complex.exp (-(1 / 2) * ↑(self.Q f f))
Instances For
The negation map on field configurations: T(ω) = -ω
Equations
Instances For
The negation map is measurable w.r.t. the cylinder σ-algebra.
theorem
MinlosAnalytic.integral_neg_invariance
[IsHilbertNuclear TestFunction]
[TopologicalSpace.SeparableSpace TestFunction]
[Nonempty TestFunction]
[IsTopologicalAddGroup TestFunction]
[ContinuousSMul ℝ TestFunction]
(C : CovarianceForm)
(μ : MeasureTheory.ProbabilityMeasure FieldConfiguration)
(h_realCF :
∀ (f : TestFunction),
∫ (ω : FieldConfiguration), Complex.exp (Complex.I * ↑(ω f)) ∂↑μ = Complex.exp (-(1 / 2) * ↑(C.Q f f)))
(f : FieldConfiguration → ℂ)
:
MeasureTheory.Integrable f ↑μ → ∫ (ω : FieldConfiguration), f ω ∂↑μ = ∫ (ω : FieldConfiguration), f (-ω) ∂↑μ
Symmetry under global sign flip induced by the real Gaussian CF. Uses Minlos uniqueness from the bochner library.
theorem
MinlosAnalytic.moment_zero_from_realCF
[IsHilbertNuclear TestFunction]
[TopologicalSpace.SeparableSpace TestFunction]
[Nonempty TestFunction]
[IsTopologicalAddGroup TestFunction]
[ContinuousSMul ℝ TestFunction]
(C : CovarianceForm)
(μ : MeasureTheory.ProbabilityMeasure FieldConfiguration)
(h_realCF :
∀ (f : TestFunction),
∫ (ω : FieldConfiguration), Complex.exp (Complex.I * ↑(ω f)) ∂↑μ = Complex.exp (-(1 / 2) * ↑(C.Q f f)))
(a : TestFunction)
(hInt1 : MeasureTheory.Integrable (fun (ω : FieldConfiguration) => ↑(ω a)) ↑μ)
:
Zero mean from the real Gaussian characteristic functional, via symmetry and L¹.