OAI/NumberTheory/Ostmann/Construction/InitialHalfListWeights.lean
5 declarations, 0 contain sorry, 0 declared as axioms
The two independent cutoff weights in an initial half-list
- def
OAI.Ostmann.initialLogSumWeightline 11 - theorem
OAI.Ostmann.initialLogSumWeight_nonnegline 15 - theorem
OAI.Ostmann.initialLogSumWeight_le_oneline 19 - theorem
OAI.Ostmann.balancedHalfMass_appendline 25Source documentation
Joining the bulk and spectator groups preserves the two weights and both balanced events under their original independent priors.
theorem balancedHalfMass_append {A : Type*} [Fintype A] (b d : ℕ) (μ : Fin b → A → ℝ) (ν : Fin d → A → ℝ) (goodB : Fin b → A → Prop) (goodD : Fin d → A → Prop) (wB : (Fin b → A) → ℝ) (wD : (Fin d → A) → ℝ) : (∑ x ∈ balancedTupleSet (b + d) (Fin.append goodB goodD), productPrior (Fin.append μ ν) x * (wB (fun i => x (i.castAdd d)) * wD (fun i => x (i.natAdd b)))) = (∑ x ∈ balancedTupleSet b goodB, productPrior μ x * wB x) * ∑ x ∈ balancedTupleSet d goodD, productPrior ν x * wD x := by …Mentions: OAI.Ostmann.balancedTupleSet, OAI.Ostmann.productPrior
- theorem
OAI.Ostmann.balancedHalfMass_log_weights_lowerline 52Source documentation
The two selected cutoffs together retain a fixed exponential amount of balanced mass. Their product is still a weight, not a conditioned law.
theorem balancedHalfMass_log_weights_lower {A : Type*} [Fintype A] (b d m : ℕ) (value : A → ℕ) (μ : Fin b → A → ℝ) (ν : Fin d → A → ℝ) (goodB : Fin b → A → Prop) (goodD : Fin d → A → Prop) (cb cd C : ℝ) (hB : Real.exp (-C * m) ≤ ∑ x ∈ balancedTupleSet b goodB, productPrior μ x * initialLogSumWeight value cb x) (hD : Real.exp (-C * m) ≤ ∑ x ∈ balancedTupleSet d goodD, productPrior ν x * initialLogSumWeight value cd x) : Real.exp (-(2 * C) * m) ≤ ∑ x ∈ balancedTupleSet (b + d) (Fin.append goodB goodD), productPrior (Fin.append μ ν) x * (initialLogSumWeight value cb (fun i => x (i.castAdd d)) * …Mentions: OAI.Ostmann.balancedTupleSet, OAI.Ostmann.productPrior, OAI.Ostmann.initialLogSumWeight