SearcharxivSearch

arXiv subjects

Akshat Mudgal

Publications and source records attributed to Akshat Mudgal.

At least 19 recordsLinked to original sources

On an asymmetric additive energy inequality

Let $d \geq 1$ be an integer, $G$ be an abelian group and $ν, w_1, \dots, w_{2d}: G \to [0, \infty)$ be functions with finite, non-empty supports. Define the generalised additive energy \[ E_{2d, ν}(w_1, \dots, w_{2d}) = \sum_{y,y' \in G}\sum_{a_1, \dots, a_{2d} \in G } w_1(a_1) \dots w_{2d}(a_{2d}) ν(y) ν(y') 1_{\sum_{i=1}^d (a_i - a_{i+d}) = y-y'} .\] Moreover, for every $1 \leq i \leq 2d$, let $E_{2d, ν}(w_i) = E_{2d, ν}(w_i, \dots, w_i)$. A standard Fourier analytic argument delivers the estimate \[ E_{2d,ν}(w_1, \dots, w_{2d}) \leq \prod_{1 \leq i \leq 2d} E_{2d, ν}(w_i)^{1/2d}.\] In this note, we present a purely combinatorial proof of the above inequality. In particular, our proof does not use any Fourier or spectral analysis and relies on repeated applications of Cauchy--Schwarz inequality combined with a discrete convexity extension type argument. We also record a variation of this upper bound in the non-abelian setting via spectral inequalities following work of Hatami on graph norms, as well as a relevant sumset analogue obtained via iterative applications of the Plünnecke--Ruzsa inequality.

math.NT

Counting $2 \times 2$ integer matrices with a given determinant

Given positive integers $h, N$ satisfying $1 \leqslant h \leqslant 2N^2$, we define $T(h,N)$ to be the number of $2\times 2$ integer matrices with determinant equal to $h$ whose entries lie in $[-N,N]$. Our main result states that for any $\varepsilon >0$, one has \[ T(h,N) = \frac{16}{ζ(2)} N^2 \bigg( \sum_{d |h} \frac{1}{d} \bigg) + O_{\varepsilon}(N^{\varepsilon} (N+ h)).\] This quantitatively improves upon recent work of Afifurrahman and Ganguly--Guria, and delivers square-root cancellation estimates when $h \leq N$. We further show that when $h$ is large, the error term is of approximately the correct order.

math.NT

Counting solutions to the quadratic determinant equation

Given $h, N \in \mathbb{N}$ satisfying $1 \leqslant h \leqslant N^2$, we prove an asymptotic formula for the number of solutions to the equation $x_1 x_2 - x_3 x_4 = h$ with $x_1, \ldots, x_4 \in [-N,N] \cap \mathbb{Z}$. We use a combination of combinatorial and analytic arguments in physical space along with bounds for Kloosterman sums. Our main result concerns the case when $h = N^2 + O(N)$, wherein we obtain square-root cancellation error terms by bypassing Kloosterman sum bounds and exploiting an additional symmetry available in this setting via Ramanujan sums. This confirms a speculation of Dhanda-Haynes-Prasala in a very general form.

math.NT

A structure theorem for sets with doubling $4+δ$

We prove a structural result for sets of integers with doubling at most $4 + δ$, with $δ>0$ sufficiently small. This generalises earlier work of Eberhard--Green--Manners which dealt with sets of integers with doubling strictly less than $4$, and makes progress towards a question of Green.

math.NT

Uniform sum-product phenomenon for algebraic groups and Bremner's conjecture

In this paper we combine methods from additive combinatorics and Diophantine geometry to study the generalised sum-product phenomenon in algebraic groups. As an application of this circle of ideas, we resolve a conjecture of Bremner on arithmetic progressions in coordinates of elliptic curves, along with various other generalisations studied in the literature. We also prove a uniform Bourgain--Chang-type sum-product estimate for general $1$-dimensional algebraic groups $G$ over $\mathbb{C}$. Using these ideas, we provide an alternative solution to a problem of Bays--Breuillard. Furthermore, we show an Elekes--Szabó type result in the same setting for sets with small doubling, improving upon an earlier result of Bays--Breuillard when $G$ is not $\mathbb{G}_a$. Our power saving here can be shown to be quantitatively optimal. We use a combination of deep, classical results in Diophantine geometry due to David--Philippon, Laurent and Evertse--Schmidt--Schlickewei along with the recent breakthrough work on the weak Polynomial Freiman--Ruzsa conjecture over integers due to Gowers--Green--Manners--Tao.

math.NT

Generalised Fermat equations in dense variables over finite fields and rings

Let $A$ be a sufficiently dense subset of a finite field $\mathbb F_q$ or a finite, cyclic ring $\mathbb Z/ N\mathbb Z$. Assuming that $q$ and $N$ have no small prime divisors, we show that generalised Fermat equations have the expected number of solutions over $A$. We further show that our density threshold is optimal. Our proofs involve average Fourier decay for Bohr sets, mixed character sum bounds, equidistribution of polynomial sequences, popular Cauchy--Davenport lemmas, and a regularity-type lemma due to Semchankau.

math.NT

On exponential Freiman dimension

The exponential Freiman dimension of a finite set $A \subset \mathbb{R}^{m}$, introduced by Green and Tao in 2006, represents the largest positive integer $d$ for which $A$ contains the vertices of a non-degenerate $d$-dimensional parallelepiped. For every $d \geq 1$, we precisely determine the largest constant $C_{d}>0$ (exponential in $d$) for which $$|A+A| \geq C_{d}|A| - O_{d}(1)$$ holds for all sets $A$ with exponential Freiman dimension $d$.

math.CO

On commuting integer matrices

Given $d, N \in \mathbb{N}$, we define $\mathfrak{C}_d(N)$ to be the number of pairs of $d\times d$ matrices $A,B$ with entries in $[-N,N] \cap \mathbb{Z}$ such that $AB = BA$. We prove that $$ N^{10} \ll \mathfrak{C}_3(N) \ll N^{10},$$ thus confirming a speculation of Browning-Sawin-Wang. We further establish that $$ \mathfrak{C}_2(N) = K(2N+1)^5 (1 + o(1)),$$ where $K>0$ is an explicit constant. Our methods are completely elementary and rely on upper bounds of the correct order for restricted divisor correlations with high uniformity.

math.NT

On commuting pairs in arbitrary sets of 2x2 matrices

Let $\textrm{Mat}_2(\mathbb{R})$ be the set of $2 \times 2$ matrices with real entries. For any $\varepsilon>0$ and any finitely--supported probability measure $μ$ on $\textrm{Mat}_2(\mathbb{R})$, we prove that either \[ T(μ) = \sum_{X, Y \in {\rm supp}(μ), XY = YX} μ(X) μ(Y) < \varepsilon \] or there exists some finite set ${S}$ contained in a $2$-dimensional subspace of $\textrm{Mat}_2(\mathbb{R})$ such that $μ({S}) \geq \varepsilon/8$. This is sharp up to the multiplicative constant. We prove quantitatively stronger results when \[ μ( (a_{i,j})_{1 \leq i,j \leq 2} ) = ν(a_{1,1}) \dots ν(a_{2,2}) \ \ \text{for every} \ a_{1,1}, \dots, a_{2,2} \in \mathbb{R}, \] with $ν$ being some finitely--supported probability measure on $\mathbb{R}$. For instance, when ${A} \subset \mathbb{R}$ is a generalised arithmetic progression or multiplicative progression of dimension $d$ and $ν= {1}_{A}/|{A}|$, our techniques imply that $|{A}|^{-3} \ll_d T(μ) \ll_d |{A}|^{-3}$. Our methods highlight the connections of this problem to results in incidence geometry, growth in groups phenomenon as well as Bourgain--Chang type sum-product estimates over $\mathbb{R}$. The latter includes applications of Schmidt's subspace theorem and the resolution of the weak polynomial Freiman--Ruzsa conjecture over integers.

math.NT

Brunn-Minkowski type estimates for certain discrete sumsets

Let $d,k$ be natural numbers and let $\mathcal{L}_1, \dots, \mathcal{L}_k \in \mathrm{GL}_d(\mathbb{Q})$ be linear transformations such that there are no non-trivial subspaces $U, V \subseteq \mathbb{Q}^d$ of the same dimension satisfying $\mathcal{L}_i(U) \subseteq V$ for every $1 \leq i \leq k$. For every non-empty, finite set $A \subset \mathbb{R}^d$, we prove that \[ |\mathcal{L}_1(A) + \dots + \mathcal{L}_k(A) | \geq k^d |A| - O_{d,k}(|A|^{1- δ}), \] where $δ>0$ is some absolute constant depending on $d,k$. Building on work of Conlon-Lim, we can show stronger lower bounds when $k$ is even and $\mathcal{L}_1, \dots, \mathcal{L}_k$ satisfy some further incongruence conditions, consequently resolving various cases of a conjecture of Bukh. Moreover, given any $d, k\in \mathbb{N}$ and any finite, non-empty set $A \subset \mathbb{R}^d$ not contained in a translate of some hyperplane, we prove sharp lower bounds for the cardinality of the $k$-fold sumset $kA$ in terms of $d,k$ and $|A|$. This can be seen as a $k$-fold generalisation of Freiman's lemma.

math.CO

Finding large additive and multiplicative Sidon sets in sets of integers

Given $h,g \in \mathbb{N}$, we write a set $X \subset \mathbb{Z}$ to be a $B_{h}^{+}[g]$ set if for any $n \in \mathbb{Z}$, the number of solutions to the additive equation $n = x_1 + \dots + x_h$ with $x_1, \dots, x_h \in X$ is at most $g$, where we consider two such solutions to be the same if they differ only in the ordering of the summands. We define a multiplicative $B_{h}^{\times}[g]$ set analogously. In this paper, we prove, amongst other results, that there exist absolute constants $g \in \mathbb{N}$ and $δ>0$ such that for any $h \in \mathbb{N}$ and for any finite set $A$ of integers, the largest $B_{h}^{+}[g]$ set $B$ inside $A$ and the largest $B_{h}^{\times}[g]$ set $C$ inside $A$ satisfy \[ \max \{ |B| , |C| \} \gg_{h} |A|^{(1+ δ)/h }. \] In fact, when $h=2$, we may set $g = 31$, and when $h$ is sufficiently large, we may set $g = 1$ and $δ\gg (\log \log h)^{1/2 - o(1)}$. The former makes progress towards a recent conjecture of Klurman--Pohoata and quantitatively strengthens previous work of Shkredov.

math.NT

Unbounded expansion of polynomials and products

Given $d,s \in \mathbb{N}$, a finite set $A \subseteq \mathbb{Z}$ and polynomials $φ_1, \dots, φ_{s} \in \mathbb{Z}[x]$ such that $1 \leq deg φ_i \leq d$ for every $1 \leq i \leq s$, we prove that \[ |A^{(s)}| + |φ_1(A) + \dots + φ_s(A) | \gg_{s,d} |A|^{η_s} , \] for some $η_s \gg_{d} \log s / \log \log s$. Moreover if $φ_i(0) \neq 0$ for every $1 \leq i \leq s$, then \[ |A^{(s)}| + |φ_1(A) \dots φ_s(A) | \gg_{s,d} |A|^{η_s}. \] These generalise and strengthen previous results of Bourgain--Chang, Pálvölgyi--Zhelezov and Hanson--Roche-Newton--Zhelezov. We derive these estimates by proving the corresponding low-energy decompositions. The latter furnish further applications to various problems of a sum-product flavour, including questions concerning large additive and multiplicative Sidon sets in arbitrary sets of integers.

math.NT

A Quadratic Vinogradov Mean Value Theorem in Finite Fields

Let $p$ be a prime, let $s \geq 3$ be a natural number and let $A \subseteq \mathbb{F}_p$ be a non-empty set satisfying $|A| \ll p^{1/2}$. Denoting $J_s(A)$ to be the number of solutions to the system of equations \[ \sum_{i=1}^{s} (x_i - x_{i+s}) = \sum_{i=1}^{s} (x_i^2 - x_{i+s}^2) = 0, \] with $x_1, \dots, x_{2s} \in A$, our main result implies that \[ J_s(A) \ll |A|^{2s - 2 - 1/9}. \] This can be seen as a finite field analogue of the quadratic Vinogradov mean value theorem. Our techniques involve a variety of combinatorial geometric estimates, including studying incidences between cartesian products $A\times A$ and a special family of modular hyperbolae.

math.NT

An Elekes-Rónyai theorem for sets with few products

Given $d,n \in \mathbb{N}$, we write a polynomial $F \in \mathbb{C}[x_1,\dots,x_n]$ to be degenerate if there exist $P\in \mathbb{C}[y_1, \dots, y_{n-1}]$ and $m_j = x_1^{v_{j,1}}\dots x_n^{v_{j,n}}$ with $v_{j,1}, \dots, v_{j,n} \in \mathbb{Q}$, for every $1 \leq j \leq n-1$, such that $F = P(m_1, \dots, m_{n-1})$. Our main result shows that whenever $F$ is non-degenerate, then for every finite set $A\subseteq \mathbb{C}$ such that $|A\cdot A| \leq K|A|$, one has \[ |F(A, \dots, A)| \gg_{d,n} |A|^n 2^{-O_{d,n}((\log 2K)^{3 + o(1)})}. \] This is sharp up to a factor of $O_{d,n,K}(1)$ since we have the upper bound $|F(A,\dots,A)| \leq |A|^n$ and the fact that for every degenerate $F$ and finite set $A \subseteq \mathbb{C}$ with $|A\cdot A| \leq K|A|$, one has \[ |F(A,\dots,A)| \ll K^{O_F(1)}|A|^{n-1}.\] Our techniques rely on a variety of combinatorial and linear algebraic arguments combined with Freiman type inverse theorems and Schmidt's subspace theorem.

math.CO

Kemperman's inequality and Freiman's lemma via few translates

Let $G$ be a connected compact group equipped with the normalised Haar measure $μ$. Our first result shows that given $α, β>0$, there is a constant $c = c(α,β)>0$ such that for any compact sets $A,B\subseteq G$ with $ αμ(B)\geqμ(A)\geq μ(B) $ and $ μ(A)+μ(B)\leq 1-β$, there exist $b_1,\dots b_c\in B$ such that \[ μ(A\cdot \{b_1,\dots,b_c\})\geq μ(A)+μ(B).\] A special case of this, that is, when $G=\mathbb{T}^d$, confirms a recent conjecture of Bollobás, Leader and Tiba. We also prove a quantitatively stronger version of such a result in the discrete setting of $\mathbb{R}^d$. Thus, given $d \in \mathbb{N}$, we show that there exists $c = c(d) >0$ such that for any finite, non-empty set $A \subseteq \mathbb{R}^d$ which is not contained in a translate of a hyperplane, one can find $a_1, \dots, a_c \in A$ satisfying \[ |A+ \{a_1, \dots, a_c\}| \geq (d+1)|A| - O_d(1). \] The main term here is optimal and recovers the bounds given by Freiman's lemma up to the $O_d(1)$ error term.

math.CO

New lower bounds for cardinalities of higher dimensional difference sets and sumsets

Let $d \geq 4$ be a natural number and let $A$ be a finite, non-empty subset of $\mathbb{R}^d$ such that $A$ is not contained in a translate of a hyperplane. In this setting, we show that \[ |A-A| \geq \bigg(2d - 2 + \frac{1}{d-1} \bigg) |A| - O_{d}(|A|^{1- δ}), \] for some absolute constant $δ>0$ that only depends on $d$. This provides a sharp main term, consequently answering questions of Ruzsa and Stanchescu up to an $O_{d}(|A|^{1- δ})$ error term. We also prove new lower bounds for restricted type difference sets and asymmetric sumsets in $\mathbb{R}^d$.

math.CO

Diameter free estimates for the quadratic Vinogradov mean value theorem

Let $s \geq 3$ be a natural number, let $ψ(x)$ be a polynomial with real coefficients and degree $d \geq 2$, and let $A$ be some large, non-empty, finite subset of real numbers. We use $E_{s,2}(A)$ to denote the number of solutions to the system of equations \[ \sum_{i=1}^{s} (ψ(x_i) - ψ(x_{i+s}) )= \sum_{i=1}^{s} ( x_i - x_{i+s} ) = 0, \] where $x_i \in A$ for each $1 \leq i \leq 2s$. Our main result shows that \[ E_{s,2}(A) \ll_{d,s} |A|^{2s -3 + η_{s}}, \] where $η_3 = 1/2$, and $η_{s} = (1/4- 1/7246)\cdot 2^{-s + 4}$ when $s \geq 4$. The only other previously known result of this flavour is due to Bourgain and Demeter, who showed that when $ψ(x) = x^2$ and $s=3$, we have \[E_{3,2}(A) \ll_ε |A|^{3 + 1/2 + ε},\] for each $ε> 0$. Thus our main result improves upon the above estimate, while also generalising it for larger values of $s$ and more wide-ranging choices of $ψ(x)$. The novelty of our estimates is that they only depend on $d$, $s$ and $|A|$, and are independent of the diameter of $A$. Thus when $A$ is a sparse set, our results are stronger than the corresponding bounds that are provided by methods such as decoupling and efficient congruencing. Consequently, our strategy differs from these two lines of approach, and we employ techniques from incidence geometry, arithmetic combinatorics and analytic number theory. Amongst other applications, our estimates lead to stronger discrete restriction estimates for sparse sequences.

math.NT