Documentation

LeanPool.HighContrastHomogenization.Support.Book.Ch05.Theorems.Section56.VarianceEstimateQuadratic.TraceBudget

Coarse-graining support: #

Support.Book.Ch05.Theorems.Section56.VarianceEstimateQuadratic.TraceBudget

Imported from the Apache-2.0 CoarseGraining development at commit c7ddd76c08ade64fed1b8d2ca51be14dfee8deb4.

Sum of half-weighted quadratic responses for two normalizers, with reflection in the second response and subtraction of their mixed pairing.

Equations
  • One or more equations did not get rendered due to their size.
Instances For
    theorem HCPolySupport.Book.Ch05.Section56.weightedAverage_const_mul' {d : ℕ} {U : Ch02.Domain d} (Pcell : Ch02.DomainPartition U) (c : ℝ) (f : Pcell.Cell → ℝ) :
    (Pcell.weightedAverage fun (i : Pcell.Cell) => c * f i) = c * Pcell.weightedAverage f
    theorem HCPolySupport.Book.Ch05.Section56.vecDot_matVecMul_weightedMatAverage' {d : ℕ} {U : Ch02.Domain d} (Pcell : Ch02.DomainPartition U) (F : Pcell.Cell → Mat d) (x y : Vec d) :
    vecDot x (matVecMul (Pcell.weightedMatAverage F) y) = Pcell.weightedAverage fun (i : Pcell.Cell) => vecDot x (matVecMul (F i) y)
    theorem HCPolySupport.Book.Ch05.Section56.sum_weightedAverage_two_terms_sub_const {ι : Type u_1} {κ : Type u_2} [Fintype ι] [Fintype κ] (w : κ → ℝ) (hw : ∑ c : κ, w c = 1) (f g : ι → κ → ℝ) (h : ι → ℝ) :
    ∑ α : ι, (1 / 2 * ∑ c : κ, w c * f α c + 1 / 2 * ∑ c : κ, w c * g α c - h α) = ∑ c : κ, w c * ∑ α : ι, (1 / 2 * f α c + 1 / 2 * g α c - h α)