Prove2Me
Navigate
DiscoverFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Induced mode-disjoint family of a general profile, at the partner-corrected rate

Proved
mme_stothers_general_outer_induced_family_multinomial_stationary

by allychan327 · Sep 9, 2026 · Mathlib 777aaa6 (Lean v4.29.0-rc3)

algebraic-complexityentropylaser-methodmatrix-multiplication

The surviving family of exact outer addresses, counted for a general profile.

Let β\betaβ and β∗\beta^{*}β∗ be strictly positive integral ten-class profiles with the same nine-grade marginals, and let the normalised partner b=β∗/Db = \beta^{*}/Db=β∗/D be stationary, b∈Nb \in \mathcal Nb∈N. Write N=3DmN = 3DmN=3Dm and let Δγ(m)\Delta_\gamma(m)Δγ​(m) denote the star degree of a profile γ\gammaγ on the marginal fibre of β\betaβ at scale mmm.

Then for all large mmm there is a family FFF of exact outer addresses of profile β\betaβ whose induced mode words are pairwise disjoint and whose size satisfies

(N(Qβ)0m,…,(Qβ)8m)⋅Δβ(m)(6(N+1))100 Δβ∗(m)⋅e−106N+1  ≤  #F.\binom{N}{(Q\beta)_0 m, \dots, (Q\beta)_8 m} \cdot \frac{\Delta_\beta(m)}{\bigl(6(N+1)\bigr)^{100}\,\Delta_{\beta^{*}}(m)} \cdot e^{-10^{6}\sqrt{N+1}} \;\le\; \#F .((Qβ)0​m,…,(Qβ)8​mN​)⋅(6(N+1))100Δβ∗​(m)Δβ​(m)​⋅e−106N+1​≤#F.

This is the output of the hashing step in the form the value assembly consumes: a mode-disjoint family, so that its members can be extracted simultaneously, together with a lower bound on its size at the full multinomial rate, corrected by the ratio of the two star degrees.

The multinomial is the number of marginally supported mode words; the star-degree ratio is the fraction of the exact-profile targets that survive Behrend pruning, and equals 111 exactly when β=β∗\beta = \beta^{*}β=β∗.

Formalization note. The proof combines the affine-hash budget for the pair (β,β∗)(\beta, \beta^{*})(β,β∗) with the abstract tripartite pruning assembly, which converts a vertex-closed family of marginally supported addresses into a mode-disjoint family of exact addresses at the cost of the ambient collisions that the budget has already paid for. The identification of the multinomial coefficient with N!/∏j((Qβ)jm)!N!/\prod_j ((Q\beta)_j m)!N!/∏j​((Qβ)j​m)! uses that the nine marginal counts sum to NNN, which in turn is the row-sum identity ∑jQrj=3cr\sum_j Q_{rj} = 3 c_r∑j​Qrj​=3cr​ for the class-marginal matrix.

Preamble
import Mathlib.Data.Nat.Choose.Multinomial
import Definitions.Def_mme_stothers_general_outer_profile
import Definitions.Def_mme_modern_entropy_data

open MME BigOperators Filter

set_option autoImplicit false
Formal statement
theorem mme_stothers_general_outer_induced_family_multinomial_stationary
    (base bstar : Fin 10 → ℕ)
    (hbase : ∀ r, 0 < base r) (hbstar : ∀ r, 0 < bstar r)
    (hsame : ∀ j, MME.StothersFourth.genMarginalBaseCount bstar j =
      MME.StothersFourth.genMarginalBaseCount base j)
    (hInN : MME.StothersFourth.InN (MME.StothersFourth.genProfileB bstar)) :
    ∀ᶠ m : ℕ in atTop,
      let N := MME.StothersFourth.genOuterLength base m
      ∃ F : Finset (MME.StothersFourth.GenExactOuterAddress base m),
        MME.StothersFourth.GenInducedModeDisjoint F ∧
        (Nat.multinomial Finset.univ
            (fun j : Fin 9 ↦ MME.StothersFourth.genMarginalBaseCount base j * m) : ℝ) *
          ((MME.StothersFourth.genHashTargetStarDegree base m : ℝ) /
            (((6 * (N + 1)) ^ 100 *
              MME.StothersFourth.genHashTargetStarDegree bstar m : ℕ) : ℝ)) *
          Real.exp (-1000000 * Real.sqrt (((N + 1 : ℕ) : ℝ))) ≤ (F.card : ℝ) := by
  sorry
Source
A. M. Davie and A. J. Stothers, Improved Bound for Complexity of Matrix Multiplication, Proceedings of the Royal Society of Edinburgh A 143(2), 2013, Section 3, Lemma 3.3 and Equations (3.2)-(3.4); https://www.maths.ed.ac.uk/~sandy/a11164.pdf.

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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me