openAI-math-dataset / lean /OAI /Combinatorics /DiscreteConvexity /CoveringRegularization.lean
roseking's picture
Add lean source and binary files (part 43)
e63eb7f verified
Raw History Blame Contribute Delete
10.8 kB
import OAI.Combinatorics.DiscreteConvexity.CoveringBasic
import OAI.Combinatorics.DiscreteConvexity.CoveringBinomialBounds
import Mathlib.Combinatorics.Enumerative.DoubleCounting
namespace OAI
/-!
# Weighted-degree regularization
The initial family is the actual family of high weighted-degree subsets.
Double counting bounds its cost. An edge surviving this deletion has bounded
weighted degree at every one of its subsets. Grouping surviving edges by their
intersection with `B` gives the extension estimates used for pair sampling and
for residual tuples. No regularity property is assumed separately.
-/
namespace TalagrandDiscreteConvexity.Covering
open scoped BigOperators
variable {V : Type*} [Fintype V] [DecidableEq V]
/-- Count incidences between edges and subsets of cardinality at most `r`. -/
theorem incidence_count (H : Family V) (r : ℕ)
(hH : ∀ e ∈ H, e.card = r) :
(∑ I ∈ Finset.univ.filter (fun I : Finset V => I.card ≤ r),
(H.filter (fun e => I ⊆ e)).card) = H.card * 2 ^ r := by
classical
let K : Finset (Finset V) := Finset.univ.filter (fun I => I.card ≤ r)
have hbelow (e : Finset V) (he : e ∈ H) :
K.filter (fun I => I ⊆ e) = e.powerset := by
apply Finset.ext
intro I
simp only [K, Finset.mem_filter, Finset.mem_univ, true_and, Finset.mem_powerset]
constructor
· exact fun h => h.2
· intro hIe
exact ⟨(Finset.card_le_card hIe).trans_eq (hH e he), hIe⟩
calc
(∑ I ∈ K, (H.filter (fun e => I ⊆ e)).card) =
∑ e ∈ H, (K.filter (fun I => I ⊆ e)).card := by
exact Finset.sum_card_bipartiteAbove_eq_sum_card_bipartiteBelow
(fun I e : Finset V => I ⊆ e)
_ = ∑ _e ∈ H, 2 ^ r := by
apply Finset.sum_congr rfl
intro e he
rw [hbelow e he, Finset.card_powerset, hH e he]
_ = H.card * 2 ^ r := by simp
omit [Fintype V] in
theorem weightedDegree_nonneg {H : Family V} {rho : ℝ} {r : ℕ} {I : Finset V}
(hρ : 0 ≤ rho) : 0 ≤ weightedDegree H rho r I := by
classical
exact mul_nonneg (Nat.cast_nonneg _) (pow_nonneg hρ _)
/-- The weighted incidence identity is valid even at `rho = 0`. -/
theorem weighted_incidence_count (H : Family V) (rho : ℝ) (r : ℕ)
(hH : ∀ e ∈ H, e.card = r) :
(∑ I ∈ Finset.univ.filter (fun I : Finset V => I.card ≤ r),
weightedDegree H rho r I * rho ^ I.card) =
(2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r := by
classical
let K : Finset (Finset V) := Finset.univ.filter (fun I => I.card ≤ r)
have hcast : (∑ I ∈ K, ((H.filter (fun e => I ⊆ e)).card : ℝ)) =
(H.card : ℝ) * (2 : ℝ) ^ r := by
simpa only [Nat.cast_sum, Nat.cast_mul, Nat.cast_pow, Nat.cast_ofNat] using
congrArg (fun n : ℕ => (n : ℝ)) (incidence_count H r hH)
calc
(∑ I ∈ K, weightedDegree H rho r I * rho ^ I.card) =
(∑ I ∈ K, ((H.filter (fun e => I ⊆ e)).card : ℝ)) * rho ^ r := by
rw [Finset.sum_mul]
apply Finset.sum_congr rfl
intro I hI
have hIr : I.card ≤ r := (Finset.mem_filter.mp hI).2
rw [weightedDegree, mul_assoc, ← pow_add, Nat.sub_add_cancel hIr]
_ = ((H.card : ℝ) * (2 : ℝ) ^ r) * rho ^ r := by rw [hcast]
_ = (2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r := by
rw [mul_comm (H.card : ℝ) ((2 : ℝ) ^ r)]
/-- The initial generators cost at most their weighted-incidence budget divided
by the high-degree threshold. -/
theorem initial_cost_le {H : Family V} {rho T : ℝ} {r : ℕ}
(hρ : 0 ≤ rho) (hT : 0 < T) (hH : ∀ e ∈ H, e.card = r) :
cost rho (initialFamily H rho r T) ≤
T⁻¹ * (2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r := by
classical
let K : Finset (Finset V) := Finset.univ.filter (fun I => I.card ≤ r)
have hsub : initialFamily H rho r T ⊆ K := by
intro I hI
exact Finset.mem_filter.mpr ⟨Finset.mem_univ _, (mem_initialFamily.mp hI).1⟩
have hbudget : T * cost rho (initialFamily H rho r T) ≤
(2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r := by
calc
T * cost rho (initialFamily H rho r T) =
∑ I ∈ initialFamily H rho r T, T * rho ^ I.card := by
rw [cost, Finset.mul_sum]
_ ≤ ∑ I ∈ initialFamily H rho r T,
weightedDegree H rho r I * rho ^ I.card := by
apply Finset.sum_le_sum
intro I hI
exact mul_le_mul_of_nonneg_right (mem_initialFamily.mp hI).2.le
(pow_nonneg hρ _)
_ ≤ ∑ I ∈ K, weightedDegree H rho r I * rho ^ I.card := by
apply Finset.sum_le_sum_of_subset_of_nonneg hsub
intro I _ _
exact mul_nonneg (weightedDegree_nonneg hρ) (pow_nonneg hρ _)
_ = (2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r :=
weighted_incidence_count H rho r hH
calc
cost rho (initialFamily H rho r T) ≤
((2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r) / T := by
apply (le_div_iff₀ hT).mpr
simpa only [mul_comm] using hbudget
_ = T⁻¹ * (2 : ℝ) ^ r * (H.card : ℝ) * rho ^ r := by
rw [div_eq_mul_inv]
ac_rfl
/-- Actual survival implies the weighted-degree bound at every subset of the
surviving edge. -/
theorem regular_weightedDegree_le {H : Family V} {rho T : ℝ} {r : ℕ}
(hH : ∀ e ∈ H, e.card = r) {e I : Finset V}
(he : e ∈ regularEdges H rho r T) (hIe : I ⊆ e) :
weightedDegree H rho r I ≤ T := by
have heH : e ∈ H := (mem_regularEdges.mp he).1
have hIr : I.card ≤ r := (Finset.card_le_card hIe).trans_eq (hH e heH)
apply le_of_not_gt
intro hlarge
exact (mem_regularEdges.mp he).2 I (mem_initialFamily.mpr ⟨hIr, hlarge⟩) hIe
/-- Each nonempty intersection class contributes at most `T`, and the empty
classes contribute zero. -/
private theorem regular_intersection_sum_le {H : Family V} {rho T : ℝ} {r : ℕ}
(hρ : 0 ≤ rho) (hT : 0 ≤ T) (hH : ∀ e ∈ H, e.card = r)
(B I : Finset V) :
(∑ e ∈ (regularEdges H rho r T).filter (fun e => e ∩ B = I),
rho ^ (e \ B).card) ≤ T := by
classical
let F := (regularEdges H rho r T).filter (fun e => e ∩ B = I)
by_cases hne : F.Nonempty
· obtain ⟨e₀, he₀⟩ := hne
have he₀reg := (Finset.mem_filter.mp he₀).1
have he₀I : e₀ ∩ B = I := (Finset.mem_filter.mp he₀).2
have hIe₀ : I ⊆ e₀ := by rw [← he₀I]; exact Finset.inter_subset_left
have hdegree := regular_weightedDegree_le hH he₀reg hIe₀
have hsub : F ⊆ H.filter (fun e => I ⊆ e) := by
intro e he
obtain ⟨hereg, heI⟩ := Finset.mem_filter.mp he
refine Finset.mem_filter.mpr ⟨(mem_regularEdges.mp hereg).1, ?_⟩
rw [← heI]
exact Finset.inter_subset_left
have hcards : (F.card : ℝ) ≤ ((H.filter (fun e => I ⊆ e)).card : ℝ) :=
Nat.cast_le.mpr (Finset.card_le_card hsub)
calc
(∑ e ∈ F, rho ^ (e \ B).card) = (F.card : ℝ) * rho ^ (r - I.card) := by
calc
(∑ e ∈ F, rho ^ (e \ B).card) = ∑ _e ∈ F, rho ^ (r - I.card) := by
apply Finset.sum_congr rfl
intro e he
obtain ⟨hereg, heI⟩ := Finset.mem_filter.mp he
have hecard := hH e (mem_regularEdges.mp hereg).1
have hpartition := Finset.card_sdiff_add_card_inter e B
rw [heI, hecard] at hpartition
have hdiff : (e \ B).card = r - I.card := by omega
rw [hdiff]
_ = (F.card : ℝ) * rho ^ (r - I.card) := by simp
_ ≤ weightedDegree H rho r I :=
mul_le_mul_of_nonneg_right hcards (pow_nonneg hρ _)
_ ≤ T := hdegree
· have hF : F = ∅ := Finset.not_nonempty_iff_eq_empty.mp hne
change (∑ e ∈ F, rho ^ (e \ B).card) ≤ T
rw [hF, Finset.sum_empty]
exact hT
/-- Extension cost for actual regular edges, with the exact number of possible
intersections with `B`. -/
theorem regular_extension_le_card {H : Family V} {rho T : ℝ} {r : ℕ}
(hρ : 0 ≤ rho) (hT : 0 ≤ T) (hH : ∀ e ∈ H, e.card = r) (B : Finset V) :
(∑ e ∈ regularEdges H rho r T, rho ^ (e \ B).card) ≤
T * ((B.powerset.filter (fun I => I.card ≤ r)).card : ℝ) := by
classical
let D := B.powerset.filter (fun I => I.card ≤ r)
have hmap : ∀ e ∈ regularEdges H rho r T, e ∩ B ∈ D := by
intro e he
refine Finset.mem_filter.mpr ⟨Finset.mem_powerset.mpr Finset.inter_subset_right, ?_⟩
exact (Finset.card_le_card Finset.inter_subset_left).trans_eq
(hH e (mem_regularEdges.mp he).1)
calc
(∑ e ∈ regularEdges H rho r T, rho ^ (e \ B).card) =
∑ I ∈ D, ∑ e ∈ (regularEdges H rho r T).filter (fun e => e ∩ B = I),
rho ^ (e \ B).card :=
(Finset.sum_fiberwise_of_maps_to hmap (fun e => rho ^ (e \ B).card)).symm
_ ≤ ∑ _I ∈ D, T := by
apply Finset.sum_le_sum
intro I _
exact regular_intersection_sum_le hρ hT hH B I
_ = T * (D.card : ℝ) := by simp [mul_comm]
/-- The exact binomial extension bound from the covering lemma. -/
theorem regular_extension_le_choose {H : Family V} {rho T : ℝ} {r : ℕ}
(hρ : 0 ≤ rho) (hT : 0 ≤ T) (hH : ∀ e ∈ H, e.card = r) (B : Finset V) :
(∑ e ∈ regularEdges H rho r T, rho ^ (e \ B).card) ≤
T * (∑ i ∈ Finset.range (r + 1), (Nat.choose B.card i : ℝ)) := by
calc
_ ≤ T * ((B.powerset.filter (fun I => I.card ≤ r)).card : ℝ) :=
regular_extension_le_card hρ hT hH B
_ = _ := by rw [card_subsets_card_le, Nat.cast_sum]
/-- The polynomial version used for extending a residual tuple prefix. -/
theorem regular_extension_le_pow {H : Family V} {rho T : ℝ} {r : ℕ}
(hρ : 0 ≤ rho) (hT : 0 ≤ T) (hH : ∀ e ∈ H, e.card = r) (B : Finset V) :
(∑ e ∈ regularEdges H rho r T, rho ^ (e \ B).card) ≤
T * (1 + (B.card : ℝ)) ^ r := by
have hc : ((B.powerset.filter (fun I => I.card ≤ r)).card : ℝ) ≤
(((1 + B.card) ^ r : ℕ) : ℝ) :=
Nat.cast_le.mpr (card_subsets_card_le_pow B r)
have hc' : ((B.powerset.filter (fun I => I.card ≤ r)).card : ℝ) ≤
(1 + (B.card : ℝ)) ^ r := by
simpa only [Nat.cast_pow, Nat.cast_add, Nat.cast_one] using hc
exact (regular_extension_le_card hρ hT hH B).trans (mul_le_mul_of_nonneg_left hc' hT)
/-- An `r`-element prefix gives the sharper factor `2^r`, used for pairs. -/
theorem regular_extension_le_two_pow {H : Family V} {rho T : ℝ} {r : ℕ}
(hρ : 0 ≤ rho) (hT : 0 ≤ T) (hH : ∀ e ∈ H, e.card = r)
{B : Finset V} (hB : B.card = r) :
(∑ e ∈ regularEdges H rho r T, rho ^ (e \ B).card) ≤ (2 : ℝ) ^ r * T := by
have h := regular_extension_le_choose hρ hT hH B
rw [hB, ← Nat.cast_sum, Nat.sum_range_choose, Nat.cast_pow, Nat.cast_ofNat] at h
simpa only [mul_comm] using h
end TalagrandDiscreteConvexity.Covering
end OAI