[Paper Review] Simple proofs for Furstenberg sets over finite fields

Manik Dhar, Zeev Dvir+1 more|arXiv (Cornell University)|Sep 7, 2019

Limits and Structures in Graph Theory14 references4 citations

TL;DR

This paper presents elementary, combinatorial proofs for lower bounds on the size of $(k,m)$-Furstenberg sets over finite fields, significantly improving upon prior algebraic geometry-based results. Using a novel min-entropy formulation and recursive arguments, the authors establish a tight $\Omega(m^{n/k}/2^n)$ bound for general $k$ and $m \leq q^k$, while also proving stronger bounds for large $m$ using incidence estimates and pigeonholing.

ABSTRACT

A $(k,m)$-Furstenberg set $S \subset \mathbb{F}_q^n$ over a finite field is a set that has at least $m$ points in common with a $k$-flat in every direction. The question of determining the smallest size of such sets is a natural generalization of the finite field Kakeya problem. The only previously known bound for these sets is due to Ellenberg-Erman and requires sophisticated machinery from algebraic geometry. In this work we give new, completely elementary and simple, proofs which significantly improve the known bounds. Our main result relies on an equivalent formulation of the problem using the notion of min-entropy, which could be of independent interest.

Motivation & Objective

• To provide simpler, more elementary proofs for lower bounds on the size of $(k,m)$-Furstenberg sets in $\mathbb{F}_q^n$.
• To overcome the reliance on advanced algebraic geometry in prior work, particularly Ellenberg-Erman's bound.
• To establish tighter lower bounds for Furstenberg sets when $m < q^k$, a regime less understood before.
• To introduce and leverage a new min-entropy formulation of the problem as a key technical innovation.

Proposed method

• Reformulate the Furstenberg set problem using min-entropy to enable recursive, elementary arguments.
• Use the polynomial method as a base case for $k=1$, then extend to higher $k$ via induction and recursive decomposition.
• Apply a pigeonholing argument to show that any set of size $mq^{n-k}$ is trivially a $(k,m)$-Furstenberg set.
• Employ incidence estimates for large sets in finite fields to derive strong bounds when $m$ is large relative to $q$.
• Use $q$-binomial coefficients to count $k$-flats and control the number of incident points.
• Combine bounds on poor flats (those with few points in $S$) with averaging arguments to derive global lower bounds on $|S|$.

Experimental results

Research questions

• RQ1 Can elementary, non-algebraic-geometry methods be used to derive strong lower bounds for $(k,m)$-Furstenberg sets over finite fields?
• RQ2 What is the optimal lower bound for $K(q,n,k,m)$ when $m \leq q^k$, and can it be achieved without sophisticated machinery?
• RQ3 How does the min-entropy formulation of the problem simplify the analysis of Furstenberg sets?
• RQ4 Can stronger bounds be obtained when $m$ is large relative to $q$, especially when $k > n/2$?
• RQ5 What is the relationship between the number of $k$-flats and the minimal size of a set intersecting each in at least $m$ points?

Key findings

• The paper establishes a new lower bound $K(q,n,k,m) \geq \frac{1}{2^n} m^{n/k}$ for all $m \leq q^k$, significantly improving over the prior $C_{n,k} m^{n/k}$ bound with $C_{n,k}$ doubly exponential in $n$.
• For large $m$, specifically when $m \geq 2^{n+7-k} q \varepsilon^{-2}$, the bound $K(q,n,k,m) \geq (1-\varepsilon) m q^{n-k}$ holds, showing that the trivial construction is nearly optimal.
• When $k > n/2$, the bound $|S| \geq \left(1 - q^{n-2k} - \sqrt{q^{n-k} m^{-1}} \right) m q^{n-k}$ holds even without requiring one $k$-flat per direction.
• The authors show that when $n$ is divisible by $k$, the bound $K(q,n,k,m) \geq \frac{1}{2^{n/k}} m^{n/k}$ follows directly from the $k=1$ case via a simple recursive argument.
• The min-entropy reformulation is shown to be a powerful tool for analyzing Furstenberg sets, enabling elementary proofs where previous methods required deep algebraic geometry.
• The paper demonstrates that incidence estimates for large sets in finite fields can yield strong bounds without relying on the polynomial method, especially in the large-$m$ regime.