arXiv subjects

David Zywina

Publications and source records attributed to David Zywina.

At least 19 recordsLinked to original sources

arXiv2026

On the infinitude of elliptic curves over a number field with prescribed small rank

For any number field $K$ and integer $0\leq r \leq 4$, we prove that there are infinitely many elliptic curves over $K$ of rank $r$. Our elliptic curves are obtained by specializing well-chosen nonisotrivial elliptic curves over the function field $K(T)$. We use a result of Kai, which generalizes work of Green, Tao and Ziegler to number fields, to choose our specializations so that we have control over the bad primes and can perform a $2$-descent to compute ranks.

math.NT↗

arXiv2025

Rank one elliptic curves and rank stability

For any quadratic extension $L/K$ of number fields, we prove that there are infinitely many elliptic curves $E$ over $K$ so that the abelian groups $E(K)$ and $E(L)$ both have rank $1$. In particular, there are infinitely many elliptic curves of rank $1$ over any number field. This result generalizes theorems of Koymans-Pagano and Alpöge-Bhargava-Ho-Shnidman which were used to independently show that Hilbert's tenth problem over the ring of integers of any number field has a negative answer. Our approach differs since we are obtaining our elliptic curves by specializing a nonisotrivial rank $1$ family of elliptic curves and we compute all the ranks involved.

math.NT↗

arXiv2025

There are infinitely many elliptic curves over the rationals of rank 2

We show that there are infinitely many elliptic curves $E/\mathbb{Q}$, up to isomorphism over $\overline{\mathbb{Q}}$, for which the finitely generated group $E(\mathbb{Q})$ has rank exactly $2$. Our elliptic curves are given by explicit models and their rank is shown to be $2$ via a $2$-descent. That there are infinitely many such elliptic curves makes use of a theorem of Tao and Ziegler.

math.NT↗

arXiv2025

An elliptic surface with infinitely many fibers for which the rank does not jump

Let $E$ be a nonisotrivial elliptic curve over $\mathbb{Q}(T)$ and denote the rank of the abelian group $E(\mathbb{Q}(T))$ by $r$. For all but finitely many $t\in \mathbb{Q}$, specialization will give an elliptic curve $E_t$ over $\mathbb{Q}$ for which the abelian group $E_t(\mathbb{Q})$ has rank at least $r$. Conjecturally, the set of $t\in\mathbb{Q}$ for which $E_t(\mathbb{Q})$ has rank exactly $r$ has positive density. We produce the first known example for which $E_t(\mathbb{Q})$ has rank $r$ for infinitely many $t\in\mathbb{Q}$. For our particular $E/\mathbb{Q}(T)$ which has rank $0$, we will make use of a theorem of Green on $3$-term arithmetic progressions in the primes to produce $t\in\mathbb{Q}$ for which $E_t$ has only a few bad primes that we understand well enough to perform a $2$-descent.

math.NT↗

arXiv2025

Drinfeld modules with maximal Galois action

With a fixed prime power $q>1$, define the ring of polynomials $A=\mathbb{F}_q[t]$ and its fraction field $F=\mathbb{F}_q(t)$. For each pair $a=(a_1,a_2) \in A^2$ with $a_2$ nonzero, let $ϕ(a)\colon A\to F\{τ\}$ be the Drinfeld $A$-module of rank $2$ satisfying $t\mapsto t+a_1τ+a_2τ^2$. The Galois action on the torsion of $ϕ(a)$ gives rise to a Galois representation $ρ_{ϕ(a)}\colon \operatorname{Gal}(F^{\operatorname{sep}}/F)\to \operatorname{GL}_2(\widehat{A})$, where $\widehat{A}$ is the profinite completion of $A$. We show that the image of $ρ_{ϕ(a)}$ is large for random $a$. More precisely, for all $a\in A^2$ away from a set of density $0$, we prove that the index $[\operatorname{GL}_2(\widehat{A}):ρ_{ϕ(a)}(\operatorname{Gal}(F^{\operatorname{sep}}/F))]$ divides $q-1$ when $q>2$ and divides $4$ when $q=2$. We also show that the representation $ρ_{ϕ(a)}$ is surjective for a positive density set of $a\in A^2$.

math.NT↗

arXiv2024

Open image computations for elliptic curves over number fields

For a non-CM elliptic curve $E$ defined over a number field $K$, the Galois action on its torsion points gives rise to a Galois representation $ρ_E: Gal(\overline{K}/K)\to GL_2(\widehat{\mathbb{Z}})$ that is unique up to isomorphism. A renowned theorem of Serre says that the image of $ρ_E$ is an open, and hence finite index, subgroup of $GL_2(\widehat{\mathbb{Z}})$. In an earlier work of the author, an algorithm was given, and implemented, that computed the image of $ρ_E$ up to conjugacy in $GL_2(\widehat{\mathbb{Z}})$ in the special case $K=\mathbb{Q}$. A fundamental ingredient of this earlier work was the Kronecker-Weber theorem whose conclusion fails for number fields $K\neq \mathbb{Q}$. We shall give an overview of an analogous algorithm for a general number field and work out the required group theory. We also give some bounds on the index in Serre's theorem for a typical elliptic curve over a fixed number field.

math.NT↗

arXiv2024

Explicit open images for elliptic curves over $\mathbb{Q}$

For a non-CM elliptic curve $E$ defined over $\mathbb{Q}$, the Galois action on its torsion points gives rise to a Galois representation $ρ_E: Gal(\overline{\mathbb{Q}}/\mathbb{Q})\to GL_2(\widehat{\mathbb{Z}})$ that is unique up to isomorphism. A renowned theorem of Serre says that the image of $ρ_E$ is an open, and hence finite index, subgroup of $GL_2(\widehat{\mathbb{Z}})$. We describe an algorithm that computes the image of $ρ_E$ up to conjugacy in $GL_2(\widehat{\mathbb{Z}})$; this algorithm is practical and has been implemented. Up to a positive answer to a uniformity question of Serre and finding all the rational points on a finite number of explicit modular curves of genus at least $2$, we give a complete classification of the groups $ρ_E(Gal(\overline{\mathbb{Q}}/\mathbb{Q}))\cap SL_2(\widehat{\mathbb{Z}})$ and the indices $[GL_2(\widehat{\mathbb{Z}}):ρ_E(Gal(\overline{\mathbb{Q}}/\mathbb{Q}))]$ for non-CM elliptic curves $E/\mathbb{Q}$. Much of the paper is dedicated to the efficient computation of modular curves via modular forms expressed in terms of Eisenstein series.

math.NT↗

arXiv2024

Improved bounds for integral points on modular curves using Runge's method

Consider a modular curve $X_G$ defined over a number field $K$, where $G$ is a subgroup of $GL_2(\mathbb{Z}/N\mathbb{Z})$ with $N>2$. The curve $X_G$ comes with a morphism $j: X_G\to \mathbb{P}^1_K=\mathbb{A}^1_K \cup\{\infty\}$ to the $j$-line. For a finite set of places $S$ of $K$ that satisfies a certain condition, Runge's method shows that there are only finitely many points $P \in X_G(K)$ for which $j(P)$ lies in the ring $\mathfrak{O}_{K,S}$ of $S$-units of $K$. We prove an explicit version which shows that if $j(P)\in \mathfrak{O}_{K,S}$ for some $P\in X_G(K)$, then the absolute logarithmic height of $j(P)$ is bounded above by $N^{12} \log N$. Explicits upper bounds have already been obtained by Bilu and Parent though they are not polynomial in $N$. The modular functions needed to apply Runge's method are constructing using Eisenstein series of weight $1$.

math.NT↗

arXiv2023

Torsion bounds for a fixed abelian variety and varying number field

Let $A$ be an abelian variety defined over a number field $K$. For a finite extension $L/K$, the cardinality of the group $A(L)_{\operatorname{tors}}$ of torsion points in $A(L)$ can be bounded in terms of the degree $[L:K]$. We study the smallest real number $β_A$ such that for any finite extension $L/K$ and $\varepsilon>0$, we have $|A(L)_{\operatorname{tors}}| \leq C \cdot [L:K]^{β_A+\varepsilon}$, where the constant $C$ depends only on $A$ and $\varepsilon$ (and not $L$). Assuming the Mumford--Tate conjecture for $A$, we will show that $β_A$ agrees with the conjectured value of Hindry and Ratazzi. We also give a similar bound for the maximal order of a torsion point in $A(L)$.

math.NT↗

arXiv2022

Possible indices for the Galois image of elliptic curves over Q

For a non-CM elliptic curve $E$ over the rationals, the Galois action on its torsion points can be expressed in terms of a Galois representation $ρ_E : G \to GL_2(\hat{\mathbb{Z}})$, where $G$ is the absolute Galois group of the rationals. A well-known theorem of Serre says that the image of $ρ_E$ is open and hence has finite index in $GL_2(\hat{\mathbb{Z}})$. We will study what indices are possible assuming that we are willing to exclude a finite number of possible $j$-invariants from consideration. For example, we will show that there is a finite set $J$ of rational numbers such that if $E/\mathbb{Q}$ is a non-CM elliptic curve with $j$-invariant not in $ J$ and with surjective mod $\ell$ representations for all $\ell >37$ (which conjecturally always holds), then the index $[GL_2(\hat{\mathbb{Z}}) : ρ_E(G)]$ lies in the set \[ I:= \left\{\begin{array}{c}2, 4, 6, 8, 10, 12, 16, 20, 24, 30, 32, 36, 40, 48, 54, 60, 72, 84, 96, 108, 112,120, 144, \\192, 220, 240, 288, 336, 360, 384, 504, 576, 768, 864, 1152, 1200, 1296, 1536 \end{array}\right\}. \] Moreover, $I$ is the minimal set with this property.

math.NT↗

arXiv2021

Modular curves of prime-power level with infinitely many rational points

For each open subgroup $G$ of ${\rm GL}_2(\hat{\mathbb{Z}})$ containing $-I$ with full determinant, let $X_G/\mathbb{Q}$ denote the modular curve that loosely parametrizes elliptic curves whose Galois representation, which arises from the Galois action on its torsion points, has image contained in $G$. Up to conjugacy, we determine a complete list of the $248$ such groups $G$ of prime power level for which $X_G(\mathbb{Q})$ is infinite. For each $G$, we also construct explicit maps from each $X_G$ to the $j$-line. This list consists of $220$ modular curves of genus $0$ and $28$ modular curves of genus $1$. For each prime $\ell$, these results provide an explicit classification of the possible images of the $\ell$-adic Galois representations arising from elliptic curves over $\mathbb{Q}$ that is complete except for a finite set of exceptional $j$-invariants.

math.NT↗

arXiv2021

Computing actions on cusp forms

For positive integers $k$ and $N$, we describe how to compute the natural action of $SL_2(\mathbb{Z})$ on the space of cusp forms $S_k(Γ(N))$, where a cusp form is given by sufficiently many terms of its $q$-expansion. This will reduce to computing the action of the Atkin--Lehner operator on $S_k(Γ)$ for a congruence subgroup $Γ_1(N)\subseteq Γ\subseteq Γ_0(N)$. Our motivating application of such fundamental computations is to compute explicit models of some modular curves $X_G$.

math.NT↗

arXiv2020

Determining monodromy groups of abelian varieties

Associated to an abelian variety over a number field are several interesting and related groups: the motivic Galois group, the Mumford-Tate group, $\ell$-adic monodromy groups, and the Sato-Tate group. Assuming the Mumford-Tate conjecture, we show that from two well chosen Frobenius polynomials of our abelian variety, we can recover the identity component of these groups (or at least an inner form), up to isomorphism, along with their natural representations. We also obtain a practical probabilistic algorithm to compute these groups by considering more and more Frobenius polynomials; the groups are connected and reductive and thus can be expressed in terms of root datum. These groups are conjecturally linked with algebraic cycles and in particular we obtain a probabilistic algorithm to compute the dimension of the Hodge classes of our abelian variety for any fixed degree.

math.NT↗

arXiv2020

On the surjectivity of mod $\ell$ representations associated to elliptic curves

Let $E$ be an elliptic curve over the rationals that does not have complex multiplication. For each prime $\ell$, the action of the absolute Galois group on the $\ell$-torsion points of $E$ can be given in terms of a Galois representation $ρ_{E,\ell}\colon \operatorname{Gal}(\bar{\mathbb{Q}}/\mathbb{Q}) \to GL_2(\mathbb{F}_\ell)$. An important theorem of Serre says that $ρ_{E,\ell}$ is surjective for all sufficiently large $\ell$. In this paper, we describe a simple algorithm based on Serre's proof that can quickly determine the finite set of primes $\ell$ for which $ρ_{E,\ell}$ is not surjective. We will also give some improved bounds for Serre's theorem.

math.NT↗

arXiv2020

Galois groups arising from families with big orthogonal monodromy

We study the Galois groups of polynomials arising from a compatible family of representations with big orthogonal monodromy. We show that the Galois groups are usually as large as possible given the constraints imposed on them by a functional equation and discriminant considerations. As an application, we consider the Frobenius polynomials arising from the middle étale cohomology of hypersurfaces in $\mathbb{P}_{\mathbb{F}_q}^{2n+1}$ of degree at least $3$. We also consider the $L$-functions of quadratic twists of fixed degree of an elliptic curve over a function field $\mathbb{F}_q(t)$. To determine the typical Galois group in the elliptic curve setting requires using some known cases of the Birch and Swinnerton-Dyer conjecture. This extends and generalizes work of Chavdarov, Katz and Jouve.

math.NT↗

arXiv2019

An effective open image theorem for abelian varieties

Fix an abelian variety $A$ of dimension $g\geq 1$ defined over a number field $K$. For each prime $\ell$, the Galois action on the $\ell$-power torsion points of $A$ induces a representation $ρ_{A,\ell}\colon Gal_K \to GL_{2g}(\mathbb{Z}_\ell)$. The $\ell$-adic monodromy group of $A$ is the Zariski closure $G_{A,\ell}$ of the image of $ρ_{A,\ell}$ in $GL_{2g,\mathbb{Q}_\ell}$. The image of $ρ_{A,\ell}$ is open in $G_{A,\ell}(\mathbb{Q}_\ell)$ with respect to the $\ell$-adic topology and hence the index $[G_{A,\ell}(\mathbb{Q}_\ell)\cap GL_{2g}(\mathbb{Z}_\ell): ρ_{A,\ell}(Gal_K)]$ is finite. We prove that this index can be bounded in terms of $g$ for all $\ell$ larger then some constant depending on certain invariants of $A$.

math.NT↗

arXiv2019

Families of abelian varieties and large Galois images

Associated to an abelian variety $A$ of dimension $g$ over a number field $K$ is a Galois representation $ρ_A\colon Gal(\bar{K}/K)\to GL_{2g}(\hat{\mathbb{Z}})$. The representation $ρ_A$ encodes the Galois action on the torsion points of $A$ and its image is an interesting invariant of $A$ that contains a lot of arithmetic information. We consider abelian varieties over $K$ parametrized by the $K$-points of a nonempty open subvariety $U\subseteq \mathbb{P}^n_K$. We show that away from a set of density $0$, the image of $ρ_A$ will be very large; more precisely, it will have uniformly bounded index in a group obtained from the family of abelian varieties. This generalizes earlier results which assumed that the family of abelian varieties have "big monodromy". We also give a version for a family of abelian varieties with a more general base.

math.NT↗

arXiv2015

An explicit Jacobian of dimension 3 with maximal Galois action

We gives an explicit genus 3 curve over Q such that the Galois action on the torsion points of its Jacobian is a large as possible. That such curves exist is a consequence of a theorem of D. Zureick-Brown and the author; however, those methods do not produce explicit examples. We shall apply the general strategies of Hall and Serre in their open image theorems. We also make use of Serre's conjecture to show that the modulo ell Galois actions are irreducible. While we computationally focus on a single curve, the methods of this paper can be applied to a large family of genus 3 curves.

math.NT↗