The Multivariate Schwartz-Zippel Lemma

TL;DR

This paper generalizes the Schwartz-Zippel lemma and Nullstellensatz to multivariate polynomials over multi-grids, providing new tools for incidence geometry and algorithms for identifying special reducible polynomials.

Contribution

It introduces a multivariate Nullstellensatz and a generalized Schwartz-Zippel lemma, along with an algorithm to detect λ-reducible polynomials, extending classical results to higher dimensions.

Findings

A multivariate Nullstellensatz certifies non-zero evaluations on multi-grids.

A generalized Schwartz-Zippel lemma applies to λ-reducible polynomials.

An algorithm detects polynomials with Cartesian product hypersurfaces in their zero set.

Abstract

Motivated by applications in combinatorial geometry, we consider the following question: Let $\lambda=(\lambda_1,\lambda_2,\ldots,\lambda_m)$ be an $m$-partition of a positive integer $n$, $S_i \subseteq \mathbb{C}^{\lambda_i}$ be finite sets, and let $S:=S_1 \times S_2 \times \ldots \times S_m \subset \mathbb{C}^n$ be the multi-grid defined by $S_i$. Suppose $p$ is an $n$-variate degree $d$ polynomial. How many zeros does $p$ have on $S$? We first develop a multivariate generalization of Combinatorial Nullstellensatz that certifies existence of a point $t \in S$ so that $p(t) \neq 0$. Then we show that a natural multivariate generalization of the DeMillo-Lipton-Schwartz-Zippel lemma holds, except for a special family of polynomials that we call $\lambda$-reducible. This yields a simultaneous generalization of Szemer\'edi-Trotter theorem and Schwartz-Zippel lemma into higher dimensions, and has applications in incidence geometry. Finally, we develop a symbolic algorithm that identifies certain $\lambda$-reducible polynomials. More precisely, our symbolic algorithm detects polynomials that include a cartesian product of hypersurfaces in their zero set. It is likely that using Chow forms the algorithm can be generalized to handle arbitrary $\lambda$-reducible polynomials, which we leave as an open problem.