Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Degree bound for a projective incidence relation

Proved
APSCounting893.Main.projective_incidence_relation_degree_bound

by Shuze Chen · Nov 18, 2026 · Mathlib 0df444a (Lean v4.33.1)

algebraic-geometryapschow-formselimination-theory

Let F:Cn→CmF:\mathbb C^n\to\mathbb C^mF:Cn→Cm have image closure of dimension 1≤r<m1\le r<m1≤r<m, suppose the origin is absent from that closure, and put d=max⁡(1,deg⁡fi)d=\max(1,\deg f_i)d=max(1,degfi​). Let WWW be the affine cone closure

W={sF(x):s∈C, x∈Cn}‾Zar.W=\overline{\{sF(x):s\in\mathbb C,\ x\in\mathbb C^n\}}^{\mathrm{Zar}}.W={sF(x):s∈C, x∈Cn}​Zar.

Fix a coordinate jjj for which fj≠0f_j\ne0fj​=0. For z∈Wz\in Wz∈W and an (r+1)×m(r+1)\times m(r+1)×m matrix BBB, define

Ai,l=zjBi,l−(∑tztBi,t)1l=j.A_{i,l}=z_jB_{i,l}-\left(\sum_t z_tB_{i,t}\right)\mathbf1_{l=j}.Ai,l​=zj​Bi,l​−(t∑​zt​Bi,t​)1l=j​.

There exists a nonzero polynomial qqq in the entries of AAA satisfying

deg⁡q≤(r+1)dr,q(A)=0for all z∈W and B.\deg q\le(r+1)d^r,\qquad q(A)=0\quad\text{for all }z\in W\text{ and }B.degq≤(r+1)dr,q(A)=0for all z∈W and B.

Every row of this parametrized matrix annihilates zzz. This is the incidence-relation form of the Chow-form degree bound used to construct the output-projection certificate. The cone and the incidence matrices are actual polynomial constructions; no avoidance property is assumed.

Preamble
import Definitions.Def_APSCounting893_Main_RandomReduction
set_option autoImplicit false
open MvPolynomial APSCounting893.Main
Formal statement
theorem APSCounting893.Main.projective_incidence_relation_degree_bound
    {n m r : ℕ} (F : PolynomialMap ℂ n m)
    (hr : 0 < r) (hrm : r < m)
    (hdim : RSCapacity.algDim (imageClosure F) = ((r : ℕ∞) : WithBot ℕ∞))
    (h0 : ¬ hasAPS F) (j : Fin m) (hj : F j ≠ 0) :
    ∃ q : MvPolynomial (Fin (r+1) × Fin m) ℂ,
      q ≠ 0 ∧ q.totalDegree ≤ (r+1)*(inputDegree F)^r ∧
      ∀ z ∈ imageClosure (fun i : Fin m => MvPolynomial.X (0 : Fin (n+1)) * MvPolynomial.rename Fin.succ (F i)),
        ∀ B : Fin (r+1) × Fin m → ℂ,
          eval (fun a => z j*B a - (∑ t : Fin m, z t*B (a.1,t))*(if a.2=j then 1 else 0)) q = 0 := by sorry
Source
Jeronimo, Krick, Sabia and Sombra, The computational complexity of the Chow form, arXiv:math/0210009, Section 1.1, pp.6–7 (incidence hypersurface and degree deg(V) in each coefficient block), https://arxiv.org/pdf/math/0210009#page=6; applied to the origin projection in Guo, Saxena and Sinhababu, Theory of Computing 15(16), 2019, proof of Theorem 4.6, p.21, https://theoryofcomputing.org/articles/v015a016/v015a016.pdf; polynomial-image degree bound there is at most d^r. This is the incidence relation underlying Agarwal–Hari–Pandey–Shukla, STOC 2027 submission #893, Theorem 3.13 and Proposition 3.14, pp.13–14, https://github.com/sigact-stoc/stoc2027/releases/download/papers-1/893.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