arXiv++ Combinatorics

Let $F$ be a field of characteristic $p>2$ and $A\subset F$ have sufficiently small cardinality in terms of $p$. We improve the state of the art of a variety of sum-product type inequalities. In particular, we prove that $$ |AA|^2|A+A|^3 \gg |A|^6,\qquad |A(A+A)|\gg |A|^{3/2}. $$ We also prove several two-variable extractor estimates: ${\displaystyle |A(A+1)| \gg|A|^{9/8},}$ $$ |A+A^2|\gg |A|^{11/10},\; |A+A^3|\gg |A|^{29/28}, \; |A+1/A|\gg |A|^{31/30}.$$ Besides, we address questions of cardinalities $|A+A|$ vs $|f(A)+f(A)|$, for a polynomial $f$, where we establish the inequalities $$ \max(|A+A|,\, |A^2+A^2|)\gg |A|^{8/7}, \;\; \max(|A-A|,\, |A^3+A^3|)\gg |A|^{17/16}. $$ Szemerédi-Trotter type implications of the arithmetic estimates in question are that a Cartesian product point set $P=A\times B$ in $F^2$, of $n$ elements, with $|B|\leq |A|< p^{2/3}$ makes $O(n^{3/4}m^{2/3} + m + n)$ incidences with any set of $m$ lines. In particular, when $|A|=|B|$, there are $\ll n^{9/4}$ collinear triples of points in $P$, $\gg n^{3/2}$ distinct lines between pairs of its points, in $\gg n^{3/4}$ distinct directions. Besides, $P=A\times A$ determines $\gg n^{9/16}$ distinct pair-wise distances. These estimates are obtained on the basis of a new plane geometry interpretation of the incidence theorem between points and planes in three dimensions, which we call collisions of images.

This paper is an erratum to our paper, entitled "On an application of Guth-Katz theorem", Math. Res. Lett. 18 (2011), no. 4, 691-697. Let $F$ be the real or complex field and $ω$ a non-degenerate skew-symmetric bilinear form in the plane $F^2$. We prove that for finite a point set $P\subset F^2\setminus\{0\}$, the set $T_ω(P)$ of nonzero values of $ω$ in $P\times P$, if nonempty, has cardinality $Ω(N^{9/13}).$ A presumably near-sharp estimate $Ω(N/\log N)$ was claimed in the abovemnetioned paper over the reals for a symmetric or skew-symmetric form $ω$. However, the set-up for the proof was flawed. We discuss why we believe that justifying this claim in full strength is a major open problem. In the special case when $P=A\times A$, where $A$ is a set of at least two reals, we establish the following sum-product type estimates: $$ |AA+ AA|= Ω\left(|A|^{19/12}\right), $$ and $$|AA-AA|= Ω\left( \frac{|A|^{26/17}}{\log^{2/17}|A|}\right).$$

We prove that a finite subset $A$ of the projective line $FP^1$ over a field $F$ of positive characteristic $p$ determines $Ω(|A|^2)$ distinct cross-ratios, as long as $|A|<\sqrt{p}.$

We discuss a unified approach to a class of geometric combinatorics incidence problems in $2D$, of the Erdös distance type. The goal is obtaining the second moment estimate, that is given a finite point set $S$ and a function $f$ on $S\times S$, an upper bound on the number of solutions of $$ f(p,p') = f(q,q')\neq 0,\qquad (p,p',q,q')\in S\times S\times S\times S. \qquad(*) $$ E.g., $f$ is the Euclidean distance in the plane, sphere, or a sheet of the two-sheeted hyperboloid. Our tool is the Guth-Katz incidence theorem for lines in $\mathbb{RP}^3$, but we focus on how the original $2D$ problem is made amenable to it. This procedure was initiated by Elekes and Sharir, based on symmetry considerations. However, symmetry considerations can be bypassed or made implicit. The classical Plücker-Klein formalism for line geometry enables one to directly interpret a solution of $(*)$ as intersection of two lines in $\mathbb{RP}^3$. This allows for a very brief argument extending the Euclidean plane distance argument to the spherical and hyperbolic distances. We also find instances of the question $(*)$ without underlying symmetry group. The space of lines in the three-space, the Klein quadric $\mathcal K$, is four-dimensional. We start out with an injective map $\mathfrak F:\,S\times S\to\mathcal K$, from a pair of points in $2D$ to a line in $3D$ and seek a combinatorial problem in the form $(*)$, which can be solved by applying the Guth-Katz theorem to the set of lines in question. We identify a few new such problems and generalise the existing ones.

Let $F$ be a field with positive odd characteristic $p$. We prove a variety of new sum-product type estimates over $F$. They are derived from the theorem that the number of incidences between $m$ points and $n$ planes in the projective three-space $PG(3,F)$, with $m\geq n=O(p^2)$, is $$O( m\sqrt{n} + km ),$$ where $k$ denotes the maximum number of collinear planes. The main result is a significant improvement of the state-of-the-art sum-product inequality over fields with positive characteristic, namely that \begin{equation}\label{mres} |A\pm A|+|A\cdot A| =Ω\left(|A|^{1+\frac{1}{5}}\right), \end{equation} for any $A$ such that $|A|<p^{\frac{5}{8}}.$

We use elementary methods to prove an incidence theorem for points and spheres in $\mathbb{F}_q^n$. As an application, we show that any point set of $P\subset \mathbb{F}_q^2$ with $|P|\geq 5q$ determines a positive proportion of all circles. The latter result is an analogue of Beck's Theorem for circles which is optimal up to multiplicative constants.

We prove an incidence theorem for points and planes in the projective space $\mathbb P^3$ over any field $\mathbb F$, whose characteristic $p\neq 2.$ An incidence is viewed as an intersection along a line of a pair of two-planes from two canonical rulings of the Klein quadric. The Klein quadric can be traversed by a generic hyperplane, yielding a line-line incidence problem in a three-quadric, the Klein image of a regular line complex. This hyperplane can be chosen so that at most two lines meet. Hence, one can apply an algebraic theorem of Guth and Katz, with a constraint involving $p$ if $p>0$. This yields a bound on the number of incidences between $m$ points and $n$ planes in $\mathbb P^3$, with $m\geq n$ as $$O\left(m\sqrt{n}+ m k\right),$$ where $k$ is the maximum number of collinear planes, provided that $n=O(p^2)$ if $p>0$. Examples show that this bound cannot be improved without additional assumptions. This gives one a vehicle to establish geometric incidence estimates when $p>0$. For a non-collinear point set $S\subseteq \mathbb F^2$ and a non-degenerate symmetric or skew-symmetric bilinear form $ω$, the number of distinct values of $ω$ on pairs of points of $S$ is $Ω\left[\min\left(|S|^{\frac{2}{3}},p\right)\right]$. This is also the best known bound over $\mathbb R$, where it follows from the Szemerédi-Trotter theorem. Also, a set $S\subseteq \mathbb F^3$, not supported in a single semi-isotropic plane contains a point, from which $Ω\left[\min\left(|S|^{\frac{1}{2}},p\right)\right]$ distinct distances to other points of $S$ are attained.

There is a parallelism between growth in arithmetic combinatorics and growth in a geometric context. While, over $\mathbb{R}$ or $\mathbb{C}$, geometric statements on growth often have geometric proofs, what little is known over finite fields rests on arithmetic proofs. We discuss strategies for geometric proofs of growth over finite fields, and show that growth can be defined and proven in an abstract projective plane -- even one with weak axioms.

New lower bounds involving sum, difference, product, and ratio sets for a set $A\subset \C$ are given. The estimates involving the sum set match, up to constants, the one obtained by Solymosi for the reals and are obtained by generalising his approach to the complex plane. The bounds involving the difference set are slightly weaker. They improve on the best known ones, including the case $A\subset \R$, which also due to Solymosi, by means of combining the use of the Szemerédi-Trotter theorem with an arithmetic combinatorics technique.

Let $P$ be a set of $N$ points in the Euclidean plane, where a positive proportion of points lies off a single straight line. This note points out two facts concerning the number of equivalence classes of triangles that $P$ determines, namely that (i) $P$ determines $Ω(N^2)$ different equivalence classes of congruent triangles, and (ii) $P$ determines $Ω(\frac{N^2}{\log N})$ different equivalence classes of similar triangles. The first fact follows from the recent theorem by Guth-Katz on point-line incidences in $\R^3$. The second one, perhaps not so well known, is due to Solymosi and Tardos.

It is shown that any subset $E$ of a plane over a finite field $\F_q$, of cardinality $|E|>q$ determines not less than $\frac{q-1}{2}$ distinct areas of triangles, moreover once can find such triangles sharing a common base. It is also shown that if $|E|\geq 64q\log_2 q$, then there are more than $\frac{q}{2}$ distinct areas of triangles sharing a common vertex. The result follows from a finite field version of the Beck theorem for large subsets of $\F_q^2$ that we prove. If $|E|\geq 64q\log_2 q$, there exists a point $z\in E$, such that there are at least $\frac{q}{4}$ straight lines incident to $z$, each supporting the number of points of $E$ other than $z$ in the interval between $\frac{|E|}{2q}$ and $\frac{2|E|}{q}.$ This is proved by combining combinatorial and Fourier analytic techniques. We also discuss higher-dimensional implications of these results in light of recent developments.

Given two points $p,q$ in the real plane, the signed area of the rectangle with the diagonal $[pq]$ equals the square of the Minkowski distance between the points $p,q$. We prove that $N>1$ points in the Minkowski plane $\R^{1,1}$ generate $Ω(\frac{N}{\log{N}})$ distinct distances, or all the distances are zero. The proof follows the lines of the Elekes/Sharir/Guth/Katz approach to the Erd\H os distance problem, analysing the 3D incidence problem, arising by considering the action of the Minkowski isometry group $ISO^*(1,1)$. The signature of the metric creates an obstacle to applying the Guth/Katz incidence theorem to the 3D problem at hand, since one may encounter a high count of congruent line intervals, lying on null lines, or "light cones", all these intervals having zero Minkowski length. In terms of the Guth/Katz theorem, its condition of the non-existence of "rich planes" generally gets violated. It turns out, however, that one can efficiently identify and discount incidences, corresponding to null intervals and devise a counting strategy, where the rich planes condition happens to be just ample enough for the strategy to succeed. As a corollary we establish the following near-optimal sum-product type estimate for finite sets $A,B\subset \R$, with more than one element: $$|(A\pm{B})\cdot{(A\pm{B})}|\gg{\frac{|A||B|}{\log{|A|}+\log{|B|}}}.$$

New lower bounds involving sum, difference, product, and ratio sets for $A\subset \C$ are given.

We prove an elementary additive combinatorics inequality, which says that if $A$ is a subset of an Abelian group, which has, in some strong sense, large doubling, then the difference set A-A has a large subset, which has small doubling.

We prove that for some universal $c$, a non-collinear set of $N>\frac{1}{c}$ points in the Euclidean plane determines at least $c \frac{N}{\log N}$ distinct areas of triangles with one vertex at the origin, as well as at least $c \frac{N}{\log N}$ distinct dot products. This in particular implies a sum-product bound $$ |A\cdot A\pm A\cdot A|\geq c\frac{|A|^2}{\log |A|} $$ for a discrete $A \subset {\mathbb R}$.

This note improves the best known exponent 1/12 in the prime field sum-product inequality (for small sets) to 1/11, modulo a logarithmic factor.

Let $P = A\times A \subset \mathbb{F}_p \times \mathbb{F}_p$, $p$ a prime. Assume that $P= A\times A$ has $n$ elements, $n<p$. See $P$ as a set of points in the plane over $\mathbb{F}_p$. We show that the pairs of points in $P$ determine $\geq c n^{1 + {1/267}}$ lines, where $c$ is an absolute constant. We derive from this an incidence theorem: the number of incidences between a set of $n$ points and a set of $n$ lines in the projective plane over $\F_p$ ($n<\sqrt{p}$) is bounded by $C n^{{3/2}-{1/10678}}$, where $C$ is an absolute constant.

We extend the recent Solymosi's sum-product estimate for reals to the complex case.

In a recent paper \cite{Gl} A. Glibichuk proved that if $A,B$ are subsets of an arbitrary finite filed $\F_q$, such that $|A||B|>q$, then $16AB = \F_q$. We improve this to $10AB = \F_q.$

In recent years, sum-product estimates in Euclidean space and finite fields have been studied using a variety of combinatorial, number theoretic and analytic methods. Erdos type problems involving the distribution of distances, areas and volumes have also received much attention. In this paper we prove a relatively straightforward function version of an incidence results for points and planes previously established in \cite{HI07} and \cite{HIKR07}. As a consequence of our methods, we obtain sharp or near sharp results on the distribution of volumes determined by subsets of vector spaces over finite fields and the associated arithmetic expressions. In particular, our machinery enables us to prove that if $E \subset {\Bbb F}_q^d$, $d \ge 4$, the $d$-dimensional vector space over a finite field ${\Bbb F}_q$, of size much greater than $q^{\frac{d}{2}}$, and if $E$ is a product set, then the set of volumes of $d$-dimensional parallelepipeds determined by $E$ covers ${\Bbb F}_q$. This result is sharp as can be seen by taking $E$ to equal to $A \times A \times ... \times A$, where $A$ is a sub-field of ${\Bbb F}_q$ of size $\sqrt{q}$. In three dimensions we establish the same result if $|E| \gtrsim q^{15/8}$. We prove in three dimensions that the set of volumes covers a positive proportion of ${\Bbb F}_q$ if $|E| \ge Cq^{3/2}$. Finally we show that in three dimensions the set of volumes covers a positive proportion of ${\Bbb F}_q$ if $|E| \ge Cq^2$, without any further assumptions on $E$, which is again sharp as taking $E$ to be a 2-plane through the origin shows.