Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 3.2 (kernels at maximum support rank): if a real matrix has maximum rank among matrices with a given support TTT, then xjys=0x_jy_s=0xj​ys​=0 for x∈ker⁡Dx\in\ker Dx∈kerD, y∈ker⁡DTy\in\ker D^{\mathsf T}y∈kerDT and (s,j)∈T(s,j)\in T(s,j)∈T

Proved
OAI.SeymourSecondNeighborhood.Pruning.lemma_3_2

by SamenHossain · Oct 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

combinatoricsgraph-theoryseymour-second-neighborhood

Let I,JI,JI,J be finite index sets and T⊆I×JT\subseteq I\times JT⊆I×J. A real matrix is supported on TTT if its entries outside TTT are zero (entries in TTT may vanish). Suppose D∈RI×JD\in\mathbb R^{I\times J}D∈RI×J is supported on TTT and has maximum rank among all real matrices supported on TTT. Then for every x∈ker⁡Dx\in\ker Dx∈kerD (that is, Dx=0Dx=0Dx=0), every y∈ker⁡DTy\in\ker D^{\mathsf T}y∈kerDT (that is, DTy=0D^{\mathsf T}y=0DTy=0) and every (s,j)∈T(s,j)\in T(s,j)∈T,

xj ys=0.x_j\,y_s=0 .xj​ys​=0.

This is Lemma 3.2 of the manuscript. Proof idea: if xjys≠0x_jy_s\neq0xj​ys​=0, the perturbed matrix D′=D+esejTD'=D+e_se_j^{\mathsf T}D′=D+es​ejT​ is still supported on TTT; D′x=xjesD'x=x_je_sD′x=xj​es​ shows es∈im⁡D′e_s\in\operatorname{im}D'es​∈imD′, and Dw=D′w−wjesDw=D'w-w_je_sDw=D′w−wj​es​ shows im⁡D⊆im⁡D′\operatorname{im}D\subseteq\operatorname{im}D'imD⊆imD′; but yyy annihilates im⁡D\operatorname{im}DimD and ys≠0y_s\neq0ys​=0, so es∉im⁡De_s\notin\operatorname{im}Des​∈/imD, whence rank⁡D′>rank⁡D\operatorname{rank}D'>\operatorname{rank}DrankD′>rankD, a contradiction. The lemma is used in the proof of the pruning lemma (Pruning.lemma_3_1) to compare restrictions of kernel vectors of two matrices.

Preamble
import Mathlib
import Definitions.Def_SeymourSecondNeighborhood
import Definitions.Def_SeymourSecondNeighborhood_Image

open Finset
Formal statement
namespace OAI.SeymourSecondNeighborhood.Pruning

theorem lemma_3_2 {I J : Type*} [Fintype I] [Fintype J] [DecidableEq I] [DecidableEq J]
    (T : Finset (I × J)) (D : Matrix I J ℝ) (hD : ∀ i j, (i, j) ∉ T → D i j = 0)
    (hmax : ∀ D' : Matrix I J ℝ, (∀ i j, (i, j) ∉ T → D' i j = 0) → D'.rank ≤ D.rank)
    (x : J → ℝ) (hx : D.mulVec x = 0) (y : I → ℝ) (hy : D.transpose.mulVec y = 0)
    (s : I) (j : J) (hsj : (s, j) ∈ T) : x j * y s = 0 := by
  sorry

end OAI.SeymourSecondNeighborhood.Pruning
Source
Statement extracted from OpenAI, *A proof of Seymour's second-neighborhood conjecture*, preprint, September 23, 2026, https://github.com/openai/math/blob/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/A-proof-of-Seymours-second-neighborhood-conjecture-September-23-2026/paper.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 paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me