OAI/Combinatorics/SharpThreshold/RestrictionInfluence.lean

GitHub source file

10 declarations, 0 contain sorry, 0 declared as axioms

Finite-product averaging for pivotal influences under restrictions.

  1. def OAI.Problem313.RestrictionInfluence.weight line 16

    Source documentation

    Bernoulli weight on an arbitrary finite set of coordinates.

    def weight (p : ℝ) (x : ι → Bool) : ℝ :=
    …
  2. def OAI.Problem313.RestrictionInfluence.influence line 20

    Source documentation

    The pivotal influence, with the redundant pivotal coordinate included in the sum.

    def influence (p : ℝ) (f : (ι → Bool) → Bool) (i : ι) : ℝ := by
    …
  3. def OAI.Problem313.RestrictionInfluence.join line 28

    Source documentation

    Glue a free and an exterior configuration into a full configuration.

    def join (x : {i // P i} → Bool) (y : {i // ¬ P i} → Bool) : ι → Bool :=
    …
  4. theorem OAI.Problem313.RestrictionInfluence.join_pos line 32
  5. theorem OAI.Problem313.RestrictionInfluence.join_neg line 37
  6. theorem OAI.Problem313.RestrictionInfluence.weight_join line 43

    Source documentation

    Product weights factor across any coordinate partition.

    theorem weight_join (p : ℝ) (x : {i // P i} → Bool)
        (y : {i // ¬ P i} → Bool) :
        weight p (join P x y) = weight p x * weight p y := by
    …

    Mentions: OAI.Problem313.RestrictionInfluence.weight, OAI.Problem313.RestrictionInfluence.join

  7. theorem OAI.Problem313.RestrictionInfluence.sum_weight_join line 51

    Source documentation

    Fubini for a coordinate restriction, with the exterior sum outermost.

    theorem sum_weight_join (p : ℝ) (g : (ι → Bool) → ℝ) :
        (∑ z, weight p z * g z) =
          ∑ y : {i // ¬ P i} → Bool, weight p y *
            ∑ x : {i // P i} → Bool, weight p x * g (join P x y) := by
    …

    Mentions: OAI.Problem313.RestrictionInfluence.weight, OAI.Problem313.RestrictionInfluence.join

  8. theorem OAI.Problem313.RestrictionInfluence.join_update line 71

    Source documentation

    Updating a free coordinate commutes with gluing an exterior configuration.

    theorem join_update (x : {i // P i} → Bool) (y : {i // ¬ P i} → Bool)
        (i : {i // P i}) (b : Bool) :
        join P (Function.update x i b) y = Function.update (join P x y) i.val b := by
    …

    Mentions: OAI.Problem313.RestrictionInfluence.join

  9. theorem OAI.Problem313.RestrictionInfluence.average_influence line 87

    Source documentation

    Averaging one pivotal influence over the exterior recovers the full influence.

    theorem average_influence (p : ℝ) (f : (ι → Bool) → Bool) (i : {i // P i}) :
        (∑ y : {i // ¬ P i} → Bool,
          weight p y * influence p (fun x => f (join P x y)) i) =
            influence p f i.val := by
    …

    Mentions: OAI.Problem313.RestrictionInfluence.weight, OAI.Problem313.RestrictionInfluence.influence, OAI.Problem313.RestrictionInfluence.join

  10. theorem OAI.Problem313.RestrictionInfluence.average_totalInfluence line 106

    Source documentation

    Averaging total influence on the free coordinates gives the sum of their original influences.

    theorem average_totalInfluence (p : ℝ) (f : (ι → Bool) → Bool) :
        (∑ y : {i // ¬ P i} → Bool, weight p y *
          ∑ i : {i // P i}, influence p (fun x => f (join P x y)) i) =
            ∑ i : {i // P i}, influence p f i.val := by
    …

    Mentions: OAI.Problem313.RestrictionInfluence.weight, OAI.Problem313.RestrictionInfluence.influence, OAI.Problem313.RestrictionInfluence.join