Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

One cofinal replication scale supports all released interior extractions

Proved
mme_released_interior_simultaneous_cofinal_graded_extraction

by Robertboy18 · Sep 23, 2026 · Mathlib 777aaa6 (Lean v4.29.0-rc3)

matrix-multiplicationregional-extraction

For any positive tolerance, repair base greater than one, positive stride and size threshold, there is one even replication scale above the threshold and divisible by the stride that supports the quantitative graded extraction for every owner and interior recipe. Each step retains its own count and repair budget; their tensor restrictions are not yet combined. The full matrix exponent bound remains a separate obligation.

Preamble
import Theorems.Thm_mme_released_interior_graded_histogram_exact_step_with_repair_scale
import Definitions.Def_mme_released_interior_integer_profiles
open BigOperators MME MME.ReleasedInterior MME.RecursiveYZ MME.MoreAsymmetryExactSeed MME.CompleteSplit MME.RegionRealization MME.ProfiledCW MME.RecursiveYZ.Certificate MME.RecursiveYZ.CWCells
open scoped Classical
Formal statement
theorem mme_released_interior_simultaneous_cofinal_graded_extraction
    (stride : ℕ) (hstride : 0 < stride)
    (d : ℕ) (hd : 1 < d) (eps : ℝ) (heps : 0 < eps) (K : ℕ) :
    ∃ k : ℕ, K ≤ k ∧ 0 < k ∧ Even k ∧ stride ∣ k ∧
    ∀ (owner : Fin 6) (s : Fin 45), (seed owner s).boundary = [] →
    let n := fun r : Fin 6 => k * (regionalSize owner s) r
    let m := fun r c => k * (splitCount owner s) r c
    let mu := fun i c w => k * (integerProfile owner s) i c w
    let source : Predicate ((k * denominator ^ 4) * 4) := fun i x =>
      (∀ p : Fin (k * denominator ^ 4),
        (∑ q, (ProfiledCW.split (ell := 3) (Equiv.refl _) rfl x p q).val) = (parent s) 0 i) ∧
      ∀ w : CompleteWord 3,
        |(Fintype.card {p : Fin (k * denominator ^ 4) //
            ProfiledCW.split (ell := 3) (Equiv.refl _) rfl x p = w} : ℝ) /
            (k * denominator ^ 4 : ℕ) -
          ((((ReleasedGlobal.jointRows owner s).map
            (fun p => if ReleasedGlobal.atom p.1 i = w then p.2 else 0)).sum : ℕ) : ℝ) /
            (denominator : ℝ) ^ 4| ≤ eps
    let keep := fun (i : Fin 2) (_ : Address 4 6 (parent s) n) =>
      parentTypical (parent_total s) n m (mu (yzMode i)) eps
    let Q := commonScale 4 (loadNum (parent_total s) m d (fun i => mu (yzMode i)) keep) (loadDen m)
    ∃ (positions : Fin ((k * denominator ^ 4) * 2) ≃ Position n)
      (reference : Address 4 6 (parent s) n), reference ∈ RecursiveXHash.target m ∧
      ∃ E : ExactStep 2 ((k * denominator ^ 4) * 4) source,
        ((RecursiveXHash.target (n := n) m).card : ℝ) *
          Real.exp (-4 * Real.sqrt (Real.log Q)) / (32 * Q) ≤ E.count ∧
        E.stage.repairExponent = Nat.log d
          (∏ i : Fin 3, Nat.card (Block 2 (fullCell (parent_total s) reference)
            (fun c i => (c.2.val i).val) mu i)) + 1 ∧
        E.output = fun i x => Graded (parent_total s) i reference
          (ProfiledCW.split (ell := 2) positions (by omega) x) ∧
          Useful (fullCell (parent_total s) reference) (mu i)
            (ProfiledCW.split (ell := 2) positions (by omega) x)  := by sorry
Source
Kernel-checked regional count identities and exact extraction with empty regions.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me