The VC-dimension and point configurations in ${\Bbb F}_q^2$

Abstract

Let $X$ be a set and ${\mathcal H}$ a collection of functions from $X$ to $\{0,1\}$. We say that ${\mathcal H}$ shatters a finite set $C \subset X$ if the restriction of ${\mathcal H}$ yields every possible function from $C$ to $\{0,1\}$. The VC-dimension of ${\mathcal H}$ is the largest number $d$ such that there exists a set of size $d$ shattered by ${\mathcal H}$, and no set of size $d+1$ is shattered by ${\mathcal H}$. Vapnik and Chervonenkis introduced this idea in the early 70s in the context of learning theory, and this idea has also had a significant impact on other areas of mathematics. In this paper we study the VC-dimension of a class of functions ${\mathcal H}$ defined on ${\Bbb F}_q^d$, the $d$-dimensional vector space over the finite field with $q$ elements. Define $$ {\mathcal H}^d_t=\{h_y(x): y \in {\Bbb F}_q^d \},$$ where for $x \in {\Bbb F}_q^d$, $h_y(x)=1$ if $||x-y||=t$, and $0$ otherwise, where here, and throughout, $||x||=x_1^2+x_2^2+\dots+x_d^2$. Here $t \in {\Bbb F}_q$, $t \not=0$. Define ${\mathcal H}_t^d(E)$ the same way with respect to $E \subset {\Bbb F}_q^d$. The learning task here is to find a sphere of radius $t$ centered at some point $y \in E$ unknown to the learner. The learning process consists of taking random samples of elements of $E$ of sufficiently large size. We are going to prove that when $d=2$, and $|E| \ge Cq^{\frac{15}{8}}$, the VC-dimension of ${\mathcal H}^2_t(E)$ is equal to $3$. This leads to an intricate configuration problem which is interesting in its own right and requires a new approach.