Documentation

MIPRE.Background.LIDT.MIPStarRE.LDT.Preliminaries.Polynomials

Polynomial preliminaries #

Packages the paper's low-individual-degree polynomial class together with Schwartz-Zippel bounds as thin wrappers around Mathlib's existing multivariate polynomial theorems.

References #

@[reducible, inline]
noncomputable abbrev MIPStarRE.LDT.Preliminaries.polyFunc (m : ) (K : Type u_1) [CommSemiring K] (d : ) :

\polyfunc{m}{q}{d} from the paper's definition of low-individual-degree polynomials. This is Mathlib's MvPolynomial.restrictDegree submodule.

Equations
Instances For

    If p ∈ polyFunc m K d, then p.degreeOf i ≤ d for every variable i.

    Paper label rem:individual-degree-convention.

    Increasing the individual-degree bound enlarges the corresponding low-degree polynomial class.

    The uniform agreement probability of two polynomials on K^m, written as a finite cardinality ratio over K^m.

    Equations
    Instances For

      If every individual degree of p is at most d, then the total degree of p is at most m * d.

      theorem MIPStarRE.LDT.Preliminaries.schwartzZippel_totalDegree {m d : } {K : Type u_1} [Field K] [Fintype K] [DecidableEq K] {g h : MvPolynomial (Fin m) K} (hneq : g h) (hg : g.totalDegree d) (hh : h.totalDegree d) :

      If two polynomials on K^m have total degree at most d, then the uniform agreement probability is at most d / |K|.

      theorem MIPStarRE.LDT.Preliminaries.schwartzZippel_individualDegree {m d : } {K : Type u_1} [Field K] [Fintype K] [DecidableEq K] (g h : (polyFunc m K d)) (hneq : g h) :
      polynomialAgreementProbability m K g h ↑(m * d) / (Fintype.card K)

      Schwartz-Zippel for the paper's low-individual-degree class \polyfunc{m}{q}{d}.