Affine-section certificate preserving the polynomial image closure
ProvedAPSCounting893.Main.affine_section_nonvanishing_certificateLet be a polynomial map with image closure of dimension , and put . There is a nonzero polynomial in the coefficients of affine equations such that
Whenever rational coefficients satisfy , their affine equations define an -dimensional plane admitting a rational affine parametrization. For every affine parametrization selected by the section predicate, the restriction of has exactly the original image closure:
This is the domain-reduction certificate used in Proposition 3.14. It preserves the entire image closure and therefore preserves both its dimension and APS. The output-projection coefficients are unrestricted in this statement.
Formalization Note The output count is written as in Lean to expose its positive size. The selected rational parametrization is the same epsilon-selected object used by the mission's input-reduction definition.
import Definitions.Def_APSCounting893_Main_RandomReduction set_option autoImplicit false open MvPolynomial
theorem APSCounting893.Main.affine_section_nonvanishing_certificate {n m r : ℕ} (F : PolynomialMap ℂ n (m+1))
(hrm : r < m+1)
(hdim : RSCapacity.algDim (imageClosure F) = ((r : ℕ∞) : WithBot ℕ∞)) :
∃ q : MvPolynomial ((Fin (n-r) × Fin (n+1))) ℂ,
q ≠ 0 ∧ q.totalDegree ≤ 2*(n-r)*(inputDegree F)^(m+1) ∧
∀ s : ReductionLabels n (m+1) r → ℚ,
eval (fun a => (s (Sum.inl a) : ℂ)) q ≠ 0 →
ParametrizesSection s (selectedSectionParameter s) ∧
imageClosure ((fun i => MvPolynomial.eval₂Hom MvPolynomial.C (parameterMap (selectedSectionParameter s)) (F i))) = imageClosure F := by sorry