Number Theory [math.NT]

Counterexamples to integer-coefficient criteria for recurrence along functions from a Hardy field

math.NT

math.CO

math.DS

We give negative answers to two questions of Bergelson, Moreira, and Richter concerning recurrence along functions from a Hardy field. For the pair \(f_1(t)=t^{3/2}\) and \(f_2(t)=\lambda t^{3/2}+t\), where \(\lambda\in\mathbb R\setminus\mathbb Q\), singled out in their integer-coefficient derivative-span question, we prove that every \(F\in\nablaz(f_1,f_2)\) satisfies \(\lim_{t\to\infty}|F(t)|\in\{0,\infty\}\). Nevertheless, there is a set \(E\subset\mathbb N\) of positive natural density such that \(R_{f_1}(E)\cap R_{f_2}(E)\) is piecewise syndetic and not thick. Thus the proposed integer-coefficient replacement does not imply thickness. We further show that, even under the same full integer derivative-span condition, the common return-time set may be empty. This stronger obstruction also gives a negative answer to their question asking whether the recurrence conclusion of Theorem A follows from joint intersectivity of the integer polynomials in \(\operatorname{poly}(f_1,\ldots,f_k)\). The constructions use elementary Bohr sets.

A Degree-Two Hilbert--P\'olya Realisation by Causal Riemann-Surface Operators

math.NT

math-ph

math.MP

math.SP

We study a local Euler-factor version of the Hilbert--Polya idea at Langlands degree two. The operators are J-self-adjoint analytic pencils whose spectral parameter lives on the Riemann surface of the square-root function. The fractional kernel is the Laplace transform of a causal response, so KK analyticity forces the branch-point cover on which the RSCO pencil lives. For the N=2 canonical case, the spectral curve is the elliptic curve E_0: y^2 = x^3 + 8x (LMFDB 256b2, conductor 256) and the local Euler factors of L(E_0,s) are recovered by an explicit off-shell basepoint in the resolvent. The basepoint is real because the Hasse--Weil bound supplies the needed inequality. The formalism extends to a family of elliptic curves through a closed-form quadratic matching system, giving a local Hilbert--Polya-type realisation at Langlands degree two. A structural discriminant identity Delta_p + D_p = 4p^2 holds at the canonical parameter point (tau,delta,Delta)=(2,0,2). The paper also proves a lambda-evenness theorem, identifies the Pontryagin index in closed form, and records obstructions to descending the construction to zeta(s). For general N, the spectral curve genus g_N equals (N-1)^2. The result is a finite, local, prime-by-prime correspondence for elliptic L-functions, not a global Hilbert--Polya realisation. The J-self-adjoint constraint that makes the elliptic matching work also blocks descent to zeta(s).

On the finite transcendence of Frobenius traces for abelian varieties over $\mathbb{Q}$

math.NT

The first purpose of this paper is to give the fnite transcendence of Frobenius traces for elliptic curves over $\mathbb{Q}$ without the assumption of complex multiplication (CM). This result generalizes the previous work by Luca and Zudilin, who obtained similar transcendence results specifically for the CM case. The second purpose is to give the finite transcendence of Frobenius traces for several principally polarized abelian varieties over $\mathbb{Q}$, by using Luca--Zudilin's method.

Multiplicative independence in the sequence of $k$-generalized Pell numbers

math.NT

We study multiplicative dependence between terms of the $k$-generalized Pell sequence $(P_n^{(k)})_{n\ge 2-k}$, defined by the linear recurrence \[ P_n^{(k)} = 2P_{n-1}^{(k)} + P_{n-2}^{(k)} + \dots + P_{n-k}^{(k)}, \] with initial conditions $P_0^{(k)} = \dots = P_{-(k-2)}^{(k)} = 0$ and $P_1^{(k)} = 1$. For $k\ge 2$ we determine all pairs $(m,n)$ with $n>m\ge 0$ such that $P_n^{(k)}$ and $P_m^{(k)}$ are multiplicatively dependent. The main result states that the only solutions occur for very small $k,m,n$ (which are listed explicitly). The proof uses lower bounds for linear forms in logarithms (Matveev), the Baker-Davenport reduction algorithm, and a computational search.

Linear independence of periods related to polylogarithms

math.NT

This paper provides the first criteria for the linear independence of multiple polylogarithm values over algebraic number fields. In particular, we derive novel results regarding the linear independence of products of polylogarithms at distinct points over an algebraic number field. Our approach is based on the explicit construction of Pad\'{e}-type approximants tailored for multiple polylogarithms.

An update on the Linnik--Goldbach and Romanov problems

math.NT

We consider the Linnik--Goldbach problem of writing all large even integers as the sum of two primes and a fixed number of powers of 2. We show that, under the generalised Riemann hypothesis, one can use 6 powers of two. In addition, we update the best known bounds on Romanov's constant, showing unconditionally that more than $25\%$ of odd numbers can be written as the sum of a prime and a power of 2.

Non-admissibility of some universal supersingular representations

math.NT

math.RT

Let $K/\mathbf{Q}_p$ be an unramified extension of degree $f$ with residue field $k$. Let $\sigma$ be an irreducible representation of $\mathrm{GL}_n(k)$ over $\overline{\mathbf{F}}_p$. For $n\ge 3$, we prove that the universal supersingular representation of weight $\sigma$ is non-admissible and of infinite length when $\sigma$ is sufficiently generic and satisfies certain technical conditions. This generalizes the previous results for $n=2$ and a non-trivial finite extension $K/\mathbf{Q}_p$. Our method employs a weight cycling argument together with recent progress on the Serre weight conjectures.

A note on the $q$-adic valuation of $\sigma_k(n)$

math.NT

In this note, we obtain an exact formula for the $q$-adic valuation of $\sigma_k(n)$ where $q$ is an odd prime, allowing us to derive an explicit upper bound which is asymptotically better than the previous bound obtained by Zhao when $n$ is large and $k \geqslant q-2$. The key parts are played by the LTE lemma and the use of cyclotomic polynomials.

Decay of weighted cusp counts for congruence subgroups of $SL_2$ over number fields

math.NT

math.AG

math.GR

For congruence subgroups commensurable with $\operatorname{SL}_2$ over number fields, we study cusp counts with certain multiplicities. We prove that the ratio of the total weighted cusp count to the group index is bounded by a negative power of the norm of the congruence level. This generalizes a theorem of Cox--Parry over $\mathbb Q$, and supports the heuristic that cusp terms occurring in topological, arithmetic and representation-theoretical formulas are subleading. The proof proceeds by localizing at a prime and reducing the problem to finite quotients, where it becomes a counting problem for finite groups. The main technical part is a counting problem for subgroups of $\operatorname{SL}_2$ over finite non-reduced principal local rings, proved by an analysis reminiscent of additive combinatorics.

On some constancy of Hecke eigensystems for Drinfeld cuspforms of level $\Gamma_1(\mathfrak{n}\wp^r)$

math.NT

Let $p$ be a rational prime, let $q>1$ be a $p$-power integer, let $\mathbb{F}_q$ be the field of $q$ elements and let $A=\mathbb{F}_q[t]$ be the polynomial ring over $\mathbb{F}_q$. Let $\mathfrak{n}\in A$ be a nonzero element and let $\wp\in A$ be a monic irreducible polynomial of positive degree. Let $k\geq 2$ and $r\geq 1$ be integers. Let $S_k(\Gamma_1(\mathfrak{n}\wp^r))$ be the space of Drinfeld cuspforms of level $\Gamma_1(\mathfrak{n}\wp^r)$ and weight $k$. In this paper, we show that a Hecke eigensystem of finite $\wp$-slope appears in $S_k(\Gamma_1(\mathfrak{n}\wp^r))$ if and only if it appears in $S_k(\Gamma_1(\mathfrak{n}\wp))$.

On the Periods of Ikeda-Yamana Lift for the Unitary Group I

math.NT

Let $F$ be a totally real field and $E$ be a quadratic CM extension field of $F$. Let $n$ be a positive integer. We denote by $\mathrm{I}_n$ the lift of the Hermitian modular form of weight $(w_{\lambda})_{\lambda}$ with level 1 to unramified automorphic forms defined on the unitary group constructed by Yamana. We then express the period $\langle{\mathrm{I}_n(f), \mathrm{I}_n(f)}\rangle$ of $\mathrm{I}_n(f)$ for Hecke eigenforms $f$ in terms of special values of some types of $L$-functions of $f$. This is an extension of Katsurada's result concerning Ikeda's conjecture.

Sum of consecutive powers as a perfect power

math.NT

In this paper we study the equation $$ x^k + (x+1)^k = y^n,\quad n\geq 3, $$ when $k\equiv 2\pmod{4}$. We prove that the only solutions are for $x=0, -1$ when $6\leq k\leq 100$ or for a $k$ with odd prime factors congruent to $3\pmod{4}$. We use linear forms in logarithms, the modular method and the resolution of Thue equations.

Classicality of Hilbert modular forms

math.NT

Let $F$ be a totally real number field. We prove that a character of the spherical Hecke algebra appearing in the completed cohomology of Hilbert modular varieties is modular if the associated Galois representation is absolutely irreducible, and de Rham of regular parallel weights. As an application, we prove some new cases of the Langlands-Clozel-Fontaine-Mazur conjecture of $\mathrm{GL}_2$ over totally real fields. For the proof, we generalize the method in [Pan26], calculate geometric partial Fontaine operators, and study the cohomology of the associated Koszul-type partial de Rham complexes. The key step is the establishment of a locally analytic Jacquet-Langlands transfer, whose proof consists of several novel ingredients including a comparison of Igusa stacks for different quaternionic Shimura data constructed by [DvHKZ26], and the Grothendieck-Messing theory for locally analytic infinite level Shimura varieties established in [Jiang26a].

Three Brillhart-Lehmer-Selfridge primality proofs for Wagstaff numbers

math.NT

The Wagstaff numbers $W_p = (2^p + 1)/3$ for odd primes $p$ are the natural $+1$ companions of the Mersenne numbers. Known primality proofs for $W_p$ with $p \geq 2617$ rely on the elliptic-curve primality proving algorithm of Atkin-Morain; Chebyshev/Lucas-type tests, while available as compositeness criteria, remain conjectural on the sufficiency side. We present fully verified primality proofs of $W_{2617}$ (788 digits), $W_{10501}$ (3161 digits), and $W_{12391}$ (3730 digits), independent of ECPP and relying only on classical $N-1$ machinery. The proofs apply the Brillhart-Lehmer-Selfridge (BLS) $N-1$ criterion to the cyclotomic decomposition $2^{p-1} - 1 = \prod_{d \mid p-1} \Phi_d(2)$, harvesting factored content from the Cunningham project tables (used as evidence) and FactorDB (used only as a discovery aid, with every retrieved factor re-certified). As an independent check on the $\mathbb{Z}[\sqrt{2}]$ arithmetic implementation, the Chua $N+1$ congruence $\omega_3^{(W_p + 1)/2} \equiv -1 \pmod{W_p}$ -- the $a=3$ case of the Chua framework with $\omega_3 = 3 + 2\sqrt{2}$, a necessary condition for Wagstaff primality -- is verified at each $W_p$. BLS $N-1$ requires $p-1$ sufficiently smooth that enough cyclotomic factors $\Phi_d(2)$ are fully factored. On the factorisation data consulted (Cunningham project tables and FactorDB, January-April 2026), $p = 10501$ and $p = 12391$ are the only exponents above $2617$ in the known Wagstaff prime/probable-prime list meeting this ceiling. Every cofactor primality is certified unconditionally by APR-CL; the method is independent of the Chebyshev sufficiency conjecture, and every step is reproducible from the archived scripts.

Covering systems where the prime divisors of all moduli are only $2$, $3$, or $5$

math.NT

We try to find all quadruples of positive integers $(m,a,b,c)$ with $a \geq b \geq c$ such that there exists a distinct covering system with minimum modulus $m$ and least common multiple of the moduli $2^a 3^b 5^c$. We obtain complete description of all such quadruples when $m=2,3,4,5$, or $6$, except when $m=6$ and $b=c=1$. We also show that if the LCM of the moduli has only $2$, $3$, or $5$ as prime divisors, then $m \leq 9$ and construct a distinct covering system with $m=8$, $a=8$, $b=3$, and $c=2$. When a covering system exists for a quadruple $(m,a,b,c)$ we provide an example. Nonexistence of covering systems is established via integer programming or by using a new estimate on the density of a set covered by a system of congruences.

Linear equations in Piatetski-Shapiro primes

math.NT

math.CO

We establish discorrelation estimates between the Piatetski-Shapiro prime set \[ \mathcal{P}_{\gamma} := \{p \text{ is prime and } p = \lfloor n^{1/\gamma} \rfloor \text{ for some } n \in \mathbb{N}\} \] and arbitrary nilsequences when $\gamma \in (0,1)$ is sufficiently close to $1$. This extends earlier works which treated linear or polynomial exponential phase functions. As an application, we establish an asymptotic formula for the number of solutions in $\mathcal{P}_{\gamma}$ to any "finite-complexity" system of linear equations, including for the number of $k$-term arithmetic progressions in $\mathcal{P}_{\gamma}$ up to a threshold $N$ for any given $k \geq 3$. Furthermore, we show that there exists an absolute constant $C>0$ such that if \[ 1 - 2^{-Ck} < \gamma < 1, \] then the Piatetski-Shapiro primes $\mathcal{P}_{\gamma}$ contain infinitely many non-trivial $k$-term arithmetic progressions. This significantly improves upon the previous range of $\gamma$ obtained by Li and Pan, which is of triple exponential type.

Epstein vector zeta functions related to the ADE Lie algebras

math-ph

math.MP

math.NT

We introduce a vector-valued generalization of the Epstein zeta functions associated with the root lattices of ADE-type Lie algebras. The quadratic forms defining these lattices correspond to the Gram matrices of the simple roots. Using the discriminant group D = P/Q, we construct vector-valued theta series that realize the Weil representation of the metaplectic group Mp(2,Z). The proposed Epstein vector zeta functions are obtained as the Mellin transform of these theta series. By exploiting the equivariance properties of the theta vectors, we derive a matrix functional equation of the Riemann type. We show that the existence of this functional equation is governed by a selection rule: it holds specifically for the subspace of C-invariant vectors, where C is the central element of Mp(2,Z). Finally, we provide a complete classification of the lattices and invariant subspaces for which this matrix functional equation is satisfied.

n-ary elliptic groups, rings, and primes in arithmetic progressions

math.RA

math.NT

I introduced the notion of an elliptic group in [Elliptic groups and rings. Beitr\"age zur Algebra und Geometrie 66(2), 497-529]. It is a quasi-group based on the tangent-chord law of elliptic curves and thus, becomes an abelian group upon singling out an element. This close proximity to abelian groups is reflected in the theory, and among other things, we can define elliptic rings, which are monoidal objects in elliptic groups. An other way of expressing this is to say that they are commutative monoids with an elliptic group structure that distributes over them. In this paper, we generalise this theory from the binary elliptic group structure to the $n$-ary structure, which we call $n$-ary elliptic groups and $n$-ary elliptic rings. The latter are once again (binary) commutative monoids with an $n$-ary operation that distributes over the monoidal structure in an $n$-ary sense. The key interest of these objects for us is their arithmetic properties, which are surprisingly pleasant. The key result is that Dirichlet's famous theorem on arithmetic progressions becomes simply Euclid's theorem in these $n$-ary rings, at least for progressions of the form $an + 1$. Motivated by the hope to eventually prove this $n$-ary Euclidean theorem purely algebraically using the theory of $n$-ary rings (and thus give an alternative and purely algebraic proof of Dirichlet's theorem), we start by exploring the first arithmetic facts of these objects, including introducing the $n$-ary class group and showing that it indeed captures the unique $n$-ary factorisation. We also obtain a type of Dedekinds theorem for our main $n$-ary ring of interest: $\mathsf{nEl}(\mathbb{Z})$.

Module Lattice Security (Part III): Structured CVP Distance on the Log-Unit Lattice

cs.DS

cs.CR

math.NT

math.ST

quant-ph

stat.TH

We prove that the $L^2$ CVP distance from a random short ring element to the log-unit lattice of $\Q(\zeta_{2^k})$ converges to $\frac{\pi}{2\sqrt{6}}\sqrt{n}$ as $n=2^{k-1}\to\infty$. We then show that this target lies inside the Voronoi cell of the origin for $k\ge 4$. For the $L^\infty$ norm, the maximum over $n$ sub-Gaussian coordinates yields $O(\sqrt{\log n})$ which translates into a sub-polynomial approximation factor for the Short Generator Problem. We show a Coarse Lattice Theorem that Babai's algorithm returns zero for all structured targets, yet exactly recovers unit perturbations of arbitrary size. For module determinant ideals, we further prove the Trigamma Theorem that proves an intrinsic imbalance $\sigma_{g_0}=O(1)$ independent of the modulus $q$. Finally, combined with Parts I and II, we reduce the CDPR factor for ML-KEM from $\exp(\tO(\sqrt{n}))$ to a sub-polynomial value.

Rationality problem for norm one tori of tensor products of \'etale algebras and Hasse norm principle

math.AG

math.NT

Let $k$ be a field. Let $A=\prod_{i=1}^r K_i$ and $B=\prod_{j=1}^s E_j$ be \'etale $k$-algebras where $K_i$ and $E_j$ are finite separable field extensions of $k$ with $[K_i:k]=m_i$ and $[E_j:k]=n_j$. Let $\mathcal{T}_A=R^{(1)}_{A/k}(\mathbb{G}_m)$ be the norm one torus of the \'etale $k$-algebra $A$. We prove that if $\gcd(m_i,n_j\mid 1\leq i\leq r, 1\leq j\leq s)=1$ and $\mathcal{T}_A$ and $\mathcal{T}_B$ are stably $($resp. retract$)$ $k$-rational, then the algebraic $k$-torus $\mathcal{T}_A\otimes \mathcal{T}_B$ and the norm one torus $\mathcal{T}_{A\otimes B}$ are stably $($resp. retract$)$ $k$-rational. We then give detailed applications to the case of norm one tori of field extensions. In particular, if $k$ is a global field, then the Hasse norm principle holds for $(A\otimes B)/k$.

Explicit cost analysis of Toom-4 multiplication for incomplete NTT in lattice-based cryptography

cs.CR

cs.SC

math.NT

Polynomial multiplication is fundamental in lattice-based cryptography. While the Number Theoretic Transform (NTT) enables fast multiplication, it imposes constraints on the modulus of the coefficient field. Hafiz et al. (2025) addressed this limitation by analyzing the incomplete NTT, which combines a truncated NTT with conventional multiplication methods In this work, we revisit Toom-4 multiplication in the context of incomplete NTT. Although Toom-4 is asymptotically faster than Karatsuba, its precise cost has not been expressed in a form compatible with the incomplete NTT framework. We present a concrete Toom-4 implementation and derive explicit operation counts that separate additions/subtractions and multiplications over the coefficient field. Our analysis based on addition chains yields a simple cost model for incomplete NTT. Using this model, we analyze hybrid strategies combining Toom-4, Karatsuba, and incomplete NTT. We identify parameter ranges where Toom-4 is advantageous and validate the predicted behavior experimentally.

Iterative maps emerging from cohomological structure of primes

cond-mat.stat-mech

math.NT

Prime numbers appeared in contexts spanning statistical mechanics, quantum mechanics and dynamical systems. However, the mechanisms governing the irregularities observed in their sequence and linking them to physical systems remained unclear. Here, it is shown that prime gaps at different separation distances follow a function depending on that distance and can be described by an iterative map which predicts the primary growth of successive primes. On the other hand, the analysis of remaining fluctuations reveals the existence of a well-defined cohomological structure, where the deterministic functional relation holds for primes up to small decaying fluctuations. In consequence, the long-range correlations as well as local jumps in primes encode the underlying cohomological structure where prime numbers are states of a given system that becomes deterministic asymptotically. Remarkably, the solution to this cohomological equation turns out to be the logarithmic integral function.

Genus formulas for dormant modular curves and asymptotic behavior of their function fields

math.AG

math.NT

Towers of algebraic function fields over finite fields play a fundamental role in arithmetic geometry and coding theory. Classical examples arising from modular and Drinfeld modular curves exhibit asymptotically good behavior. In this paper, we introduce an analogous construction derived from the moduli spaces of higher-level dormant $\mathrm{PGL}_2$-opers of prescribed radii on $4$-pointed stable curves of genus $0$. These spaces, which we refer to as dormant modular curves, form projective systems under level reduction. Building on previous results in the moduli theory of dormant opers, we establish an explicit formula for computing the genera of these curves. This formula allows us to study the asymptotic behavior of the corresponding towers of function fields and to compare them with the classical modular and Drinfeld modular cases.

The 3-adic valuations of Stirling numbers of the first kind

math.NT

v2

Let $v_3$ denote the usual $3$-adic valuation, and let $s(n, k)$ be the unsigned Stirling number of the first kind. In this paper, for $a\in\{1,2\}$, we determine the values of $v_3(s(a3^n, k))$ for all $1\le k\le a3^n$. More precisely, for each admissible pair $(m, k)$, we obtain an explicit formula for $v_3(s(a3^n, a3^m-k))$. The proof combines properties of the $m$-th Stirling numbers of the first kind with a detailed analysis of the relevant $3$-adic orders. As a consequence, we prove the case $p=3$ of a conjecture of Hong and Qiu proposed in 2020. We also derive formulas near the diagonal, comparison results for the adjacent orders $a3^n$ and $a3^n+1$, sharp upper bounds for the families $v_3(s(3^n, k))$ and $v_3(s(2\cdot3^n, k))$, and partial confirmations of conjectures of Lengyel and of Leonetti and Sanna.

On the partial derivatives of Drinfeld modular forms of arbitrary rank

math.NT

v4

In this paper, we obtain an analogue of the Serre derivation acting on the product of spaces of Drinfeld modular forms which generalizes the differential operator introduced by Gekeler in the rank two case. We further introduce a finitely generated algebra $\mathcal{M}_r$ containing all the Drinfeld modular forms for the full modular group and show its stability under the partial derivatives.

Heavenly elliptic curves over quadratic fields

math.NT

v5

An abelian variety $A/K$ is heavenly at $\ell$ if the extension $K(A[\ell^\infty])/K(\mu_{\ell^{\infty}}\!)$ is both pro-$\ell$ and unramified away from $\ell$. It is known that for a fixed quadratic field $K$, the number of $K$-isomorphism classes of heavenly elliptic curves is finite, even running over all primes $\ell$. We prove a complementary result, that for a fixed prime $\ell\geq 7$, there are only finitely many such classes, even running over all quadratic fields. This naturally raises the question of whether to expect a finiteness result when both $K$ and $\ell$ are allowed to vary. We demonstrate similarities in the behavior of heavenly elliptic curves and elliptic curves with complex multiplication, in terms of their Frobenius traces modulo $\ell$. We determine the complete list of heavenly elliptic curves defined over quadratic fields with complex multiplication and with irrational $j$-invariant (up to isomorphism). We include various extensions of our results to higher degree fields and higher-dimensional abelian varieties where possible.

Machines Learn Number Fields, But How? The Case of Galois Groups

math.NT

cs.LG

v2

By applying interpretable machine learning methods such as decision trees, we study how simple models can classify the Galois groups of Galois extensions over $\mathbb{Q}$ of degrees 4, 6, 8, 9, and 10, using Dedekind zeta coefficients. Our interpretation of the machine learning results allows us to understand how the distribution of zeta coefficients depends on the Galois group, and to prove new criteria for classifying the Galois groups of these extensions. Combined with previous results, this work provides another example of a new paradigm in mathematical research driven by machine learning.

On denominators of consecutive $\operatorname{SL}(2,{\mathbb N})$-saturated Farey fractions

math.NT

v3

The sequence $({\mathscr S}_Q)_Q$ of $\operatorname{SL}(2,{\mathbb N})$-saturated Farey fractions was defined in our previous work by ${\mathscr S}_Q := \{ a/q \in {\mathbb Q} \cap (0,1]: q+a+\bar{a} \le Q\}$, where $\bar{a}$ is the multiplicative inverse of $a\pmod{q}$ in $[1,q)$. Here, we prove that the set of $Q$-scaled denominators of consecutive fractions in ${\mathscr S}_Q$ is dense in the region ${\mathcal V}:=\{ (x,y)\in [0,1]^2 : \max \{ (1-3x)/2,2x-1\} \le y \le \max \{ x,1-x\} \}$, and provide a formula for their distribution in ${\mathcal V}$ as $Q\rightarrow \infty$.

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

math.NT

math.CO

v3

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 \[ T(h,N) = \frac{16}{\zeta(2)} N^2 \bigg( \sum_{d |h} \frac{1}{d} \bigg) + O_{\varepsilon}(N^{\varepsilon} (N+ h)).\] 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.

Modularity from $q$-series

math.NT

math.CO

v5

In 1975, G. E. Andrews challenged the mathematics community to address L. Ehrenpreis' problem, which was to directly prove the modularity of the Rogers-Ramanujan $q$-series' summatory forms. This question is important because many different $q$-series appearing in combinatorics, representation theory, and physics often seem to be mysteriously modular, yet there is no general test to confirm this directly from the exotic $q$-series expressions. In this note, we answer the challenge. We use $q$-series algebra, first-order $q$-differential systems, and analytic continuation with monodromy to give a criterion that decides when such series are modular. Specifically, we establish a necessary and sufficient condition for a vector of holomorphic $q$-series on $|q|<1$ to form a vector-valued modular function without modular input, providing a clear path to modularity for strange $q$-series.

A Dynamical N\'eron--Ogg--Shafarevich Criterion via Orbital Arboreal Representations

math.NT

math.AG

math.DS

math.GR

v3

Let $K$ be a non-archimedean local field and $\varphi : \mathbb{P}^1 \to \mathbb{P}^1$ a rational endomorphism of degree $d \geq 2$ over $K$. In the tame case ($p \nmid d$), we show that strict good reduction is equivalent to the existence of a nonempty Zariski open subset $U_k \subset \mathbb{P}^1_k \setminus \mathrm{PC}(\widetilde{\varphi})$ over which the canonical residual morphism is finite \'etale of degree $d$. The criterion separates two complementary local invariants of a normalized integral lift: $\mathrm{Res}(F,G)$ controls residual degree drop, while the fiber discriminants $\mathrm{Disc}(F_{n,x})$ control \'etaleness of the residual fibers once full residual degree is ensured. Consequently, for every finite $x \in \mathcal{O}_K$ with $\bar{x} \in U_k$, the extensions $K(X_n(x))/K$ are unramified for all $n \geq 1$. We introduce the orbital preimage tree $T_{O^+(x)} = \varinjlim_n X_\infty(\varphi^n(x))$, the colimit in $G_K$-sets along the forward orbit, and the orbital arboreal Galois image $\mathcal{G}_{O^+(x)} = \mathrm{Im}(G_K \to \mathrm{Aut}(T_{O^+(x)}))$. On the forward-invariant safe locus $U_k^{\mathrm{safe}} = \bigcap_{m \geq 0} \widetilde{\varphi}^{-m}(U_k)$, strict good reduction is captured by the bijectivity of the orbital reduction map $X_n(x_m) \to \widetilde{X}_n(\bar{x}_m)$. This canonical orbit-invariant framework connects with arboreal Galois representations (Boston-Jones, Jones, and others) and yields pointwise and orbit-level reformulations. Explicit examples over $\mathbb{Q}_p$ illustrate the criterion.

Cubic Polynomials and Sums of Two Squares

math.NT

v3

We establish a lower bound for the frequency with which an irreducible monic cubic polynomial with negative discriminant can be expressed as a sum of two squares ($\square_{2}$). This provides a quantitative answer to a question posed by Grechuk (2021) concerning the infinitude of such values. Our proof relies on a two-dimensional unit argument and the arithmetic of degree six number fields. For example, we show that if $h \equiv 2 \pmod{4}$, then \begin{align*} \# \{n : n^3+h \in \square_{2}, \ 1 \leq n \leq x \} \gg x^{1/3-o(1)}. \end{align*} These arguments may be generalised to study the representation of irreducible monic cubic polynomials by the quadratic form $x^2+ny^2$, where $n \in \mathbb{N}$.

Prime detecting quasi-modular forms in higher level

math.NT

v2

In a previous work, the authors resolved a conjecture about the structure of prime-detecting quasi-modular forms by studying sign changes occurring in quasi-modular cusp forms. In this paper, we extend the considerations to prime-detecting quasi-modular forms of higher level, in particular describing the structure of the space of quasi-modular forms that detect primes in various arithmetic progressions. We also provide an ``analytic'' proof of the level one case.

Cross representations of additive complements of $r$-th powers

math.NT

v5

Let $\mathbb{N}$ be the set of natural numbers and $\mathcal{S}_r=\big\{1^r, 2^r, 3^r,\cdots\big\}$ the set of $r$-th powers, where $r\ge 2$ is a natural number. Let $\mathcal{W}_r$ be an additive complement of $\mathcal{S}_r$ and $$ f_r(n)=\#\big\{(w,m^r)\in \mathcal{W}\times \mathcal{S}_r: n=w+m^r\big\}. $$ Motivated by a 1993 conjecture of Cilleruelo, we show that $$ \sum_{n\le N}f_r(n)-N\gg_r N^{1-\frac{1}{r}}. $$ Previously, the bound was only proved for $r=2$. In the case $r=2$, the lower bound above can be made more explicit as $$ \sum_{n\le N}f_2(n)-N\gg N^{1/2}(\log N)^{\delta} $$ for some absolute constant $\delta>0$, which improves a $\log$ factor upon a recent result of Ding, Sun, Wang and Xia.

On Periods and $L$-functions for $\mathbf{GL}_4 \times \mathbf{GL}_2$

math.NT

math.RT

v2

We give a new integral representation of the $\wedge^2 \otimes \mathrm{std}_2$ $L$-function of generic cusp forms on $\mathbf{GL}_4 \times \mathbf{GL}_2$ and $\mathbf{GU}_{2,2}\times \mathbf{GL}_2$. In the former case, we use it to prove a relation between its central $L$-value and the generalized Shalika period. Exploiting the theta correspondence for $(\mathbf{GL}_4,\mathbf{GL}_4)$, we further establish a relation between the central value of the $L$-function attached to the strongly tempered spherical pair $(\mathbf{GL}_4 \times \mathbf{GL}_2,\mathbf{GL}_2 \times \mathbf{GL}_2)$ and its corresponding period. In the case of cusp forms on $\mathbf{GL}_4 \times \mathbf{GL}_2$ that are unramified everywhere, our formulas give new evidence towards conjectures of Wan-Zhang and of Gan-Gross-Prasad for $\mathbf{GSpin}_6 \times \mathbf{GSpin}_3$.

Minimal Weierstrass models and regular models of hyperelliptic curves

math.NT

v2

Let $C$ be a hyperelliptic curve of genus $g\ge 2$ over a discrete valuation field $K$ with perfect residue field. We study the minimal Weierstrass models of $C$. When there is more than one such model, we find interesting properties on the minimal regular model and the canonical model of $C$. For curves of genus $2$, we characterize the existence of the stable reduction in terms of the minimal Weierstrass models. When there is more than one such model, we can compute the Euler factor of $\mathrm{Jac}(C)$ and a volume form of the N\'eron model of $\mathrm{Jac}(C)$, using two specific minimal Weierstrass models.

Multiple Gauss sums

math.NT

v2

A multiple Gauss sum is a complete multiple exponential sum twisted by Dirichlet characters. We prove a new bound for multiple Gauss sums and, as an application, improve previous results in the Birch--Goldbach problem. Let $F_1, \ldots, F_R \in \mathbb{Z}[x_1, \ldots, x_s]$ be forms with differing degrees, with $D$ being the highest degree, and let $\boldsymbol{F} = (F_1, \ldots, F_R)$ be nonsingular. We prove that the system $\boldsymbol{F}(\boldsymbol{x})=\mathbf{0}$ is solvable in primes provided that $s \geq D^2 4^{D+2} R^5$.

A note on Bremner's conjecture and uniformity

math.NT

v2

In 1998, Bremner conjectured that elliptic curves over the rationals having long sequences of distinct rational points whose $x$-coordinates are in arithmetic progression, have large rank. This was proved some years ago in a strong form as a consequence of previous work by the authors, by a combination of Nevanlinna theory and the uniform Mordell--Lang theorem of Gao--Ge--K\"uhne. Thus, if the ranks of elliptic curves over the rationals are uniformly bounded, then so are the lengths of the aforementioned arithmetic progressions. In this note we give a much more direct proof of this last statement, using the height-uniform Mordell theorem of Dimitrov--Gao--Habegger. The method is flexible and we give a new application of these ideas to $x$-coordinates in finitely generated multiplicative groups and geometric progressions; connections to a possible semiabelian uniform Mordell--Lang are also discussed.

Low-Lying Zeros on the Critical Line for Families of Dirichlet $L$-Functions

math.NT

v2

In this paper, we establish a new lower bound for the number of low-lying zeros of Dirichlet $L$-functions $L(s, \chi)$ on the critical line within extremely short intervals. Specifically, for a sufficiently large prime $P$ and real number $T \in [a_1/\sqrt{\log P}, 1]$, we prove that the sum of the number of zeros on the critical line $N_0(T, \chi)$ over characters $\chi \bmod P$ satisfies $$ \sum_{\chi \bmod P} N_0(T, \chi) \gg T^2 P\sqrt{\log P} .$$ Traditional approaches encounter significant technical barriers in this short-interval regime. The Levinson method fails due to its own inherent limitations in handling such restricted intervals , while standard applications of the Selberg mollifier are hindered by the emergence of complex, inseparable cross-terms that are difficult to evaluate. To overcome these obstacles, we introduce a novel analytic framework utilizing high-dimensional Mellin transforms. This approach systematically manages the multi-variable series generated by the mollifier calculations. By explicitly resolving these cross-term obstructions, we extract the localized lower bound, providing a robust method that circumvents the short-interval bottleneck and offers potential applicability to the zero statistics of higher-rank $L$-function families.

Advances in Factoring and Primality Testing: From Classical to Quantum Algorithms

cs.CR

math.NT

v2

Many modern asymmetric encryption methods rely on prime numbers, as they have distinctive properties. For instance, the security of RSA cryptosystem relies on the computational difficulty of factoring a large composite number in its prime factors, a problem that remains challenging for classical computers but potentially solvable using quantum algorithms. On the other hand, generating large prime numbers is also challenging due to their irregular distribution among integers, necessitating the use of primality testing algorithms to verify candidate primes. In this paper, we intensively review and classify various classical and quantum algorithms for factorization and primality testing, highlighting their advantages, limitations, speed/accuracy tradeoffs, time complexities, along with a brief summary. Furthermore, we apply and compare these algorithms to gain practical insights and conduct a comprehensive performance comparison. The insights from this paper show that while quantum factoring algorithms, particularly Shor's algorithm and its refinements, have introduced significant advancements over their classical counterparts, quantum primality testing algorithms have not demonstrated comparable advantages.

Projective and affine structures in positive characteristic I: Chern class formulas and Characterizations of projective spaces

math.AG

math.NT

v3

This paper aims to develop a theory of projective and affine structures on higher-dimensional varieties in positive characteristic. This theory deals with Frobenius-projective and Frobenius-affine structures, which have been previously investigated in the case where the underlying space is a curve. We first provide a description of such structures in terms of Berthelot's higher-level differential operators. That description leads us to obtain a positive characteristic version of Gunning's formulas, which give necessary conditions on Chern classes for the existence of Frobenius-projective and Frobenius-affine structures, respectively. Finally, we establish some characterizations of projective spaces using Frobenius-projective structures.

On the local Langlands conjectures for disconnected groups

math.RT

math.NT

v2

We extend the local Langlands conjectures to a certain class of disconnected groups, allowing non-abelian component groups, and recast in this language some aspects of twisted endoscopy. We further introduce normalized twisted transfer factors and a normalized correspondence between an $L$-packet for a disconnected group and the set of representations of the centralizer groups of its Langlands parameter. We prove the first instance of this conjecture, in which the identity component of the (possibly non-abelian) disconnected group is a torus.

Machine Learnability as a Measure of Order in Aperiodic Sequences

cs.LG

math.NT

v3

Research on the distribution of prime numbers has revealed a dual character: deterministic in definition yet exhibiting statistical behavior reminiscent of random processes. In this paper we show that it is possible to use an image-focused machine learning model to measure the comparative regularity of prime number fields at specific regions of an Ulam spiral. Specifically, we demonstrate that in pure accuracy terms, models trained on blocks extracted from regions of the spiral in the vicinity of 500m outperform models trained on blocks extracted from the region representing integers lower than 25m. This implies existence of more easily learnable order in the former region than in the latter. Moreover, a detailed breakdown of precision and recall scores seem to imply that the model is favouring a different approach to classification in different regions of the spiral, focusing more on identifying prime patterns at lower numbers and more on eliminating composites at higher numbers. This aligns with number theory conjectures suggesting that at higher orders of magnitude we should see diminishing noise in prime number distributions, with averages (density, AP equidistribution) coming to dominate, while local randomness regularises after scaling by log x. Taken together, these findings point toward an interesting possibility: that machine learning can serve as a new experimental instrument for number theory. Notably, the method shows potential 1 for investigating the patterns in strong and weak primes for cryptographic purposes.

A construction of single-valued elliptic polylogarithms

hep-th

math-ph

math.AG

math.MP

math.NT

v2

We establish a general construction of single-valued elliptic polylogarithms as functions on the once-punctured elliptic curve. Our formalism is an extension of Brown's construction of genus-zero single-valued polylogarithms to the elliptic curve: the condition of trivial monodromy for solutions to the Knizhnik-Zamolodchikov-Bernard equation is expressed in terms of elliptic associators and involves two representations of a two-letter alphabet. Our elliptic single-valued condition reduces to Brown's genus-zero condition upon degeneration of the torus. We provide several examples for our construction, including the elliptic Bloch-Wigner dilogarithm.

Noncommutative Wilczynski Invariants, and Modular Differential Equations

math.AG

math.NT

v2

We develop a noncommutative invariant theory for ordinary linear differential operators on Riemann surfaces. For a monic binomially normalized operator $L=\sum_{k=0}^n {n\choose k}a_kD^{\,n-k}$, $a_0=1$, with coefficients in an associative differential algebra, we construct universal gauge-covariant coefficients $I_m(L)$. After correcting their reparametrization anomalies, we obtain Wilczy\'nski currents $W_m(L)$, which transform as genuine $m$-differentials. The construction is algebraic, finite-layered, and valid over noncommutative coefficient algebras; in the commutative scalar case it recovers the classical Wilczy\'nski invariants. We globalize the theory using jet bundles and infinitesimal neighborhoods of the diagonal. The natural global objects are $A$-linear opers, where $A$ is a sheaf of associative algebras with a compatible connection. In this setting $P=I_2/(n+1)$ is an $A_{\mathrm{ad}}$-valued projective connection, while $W_m$, $m\ge 3$, are global $A_{\mathrm{ad}}$-valued differentials; scalar invariants are obtained from traces, characteristic coefficients, and cyclic trace polynomials. As applications, we discuss projective connections, symmetric powers, fanning curves in Grassmannians, Calabi--Yau Picard--Fuchs equations, weak scalar and matrix-valued $W_2$-structures from Hodge subvariations, and modular differential equations. In the modular setting, the currents become modular forms, and the first coefficient gives the modular connection underlying the Serre derivative. We also extend the formalism to Siegel space using central Siegel modular connections and the associated equivariant differential algebra.

Certified Finite-State Induction for a Perturbed Hofstadter Recursion

math.CO

math.NT

v2

We study the parity-perturbed Hofstadter-type recursion $$ Q(1)=Q(2)=1,\qquad Q(n)=Q(n-Q(n-1))+Q(n-Q(n-2))+(-1)^n . $$ We prove, by computer-certified finite-state induction, that this recursion is well-defined for all $(n\ge 1)$. The proof extracts a finite symbolic recursive model from a directly verified initial trace and then verifies an exported machine-readable certificate by independent checkers. The certificate consists of finite symbolic word systems, radius-$(R)$ contexts, context-extension records, symbolic realizations, and arithmetic recurrence records. The checkers verify symbolic closure, cycle factorization, faithfulness to an independently recomputed trace, arithmetic correctness, parity consistency, and strict backwardness of all certified recursive dependencies. The length of the computed trace is not used as evidence for global well-definedness. Instead, exhaustiveness is certified over the declared finite symbolic transition system. Together with a minimal-counterexample induction, the finite certificate rules out a first undefined recursive call.

Evaluation of two determinants involving $q$-integers

math.CO

math.NT

v2

The $q$-analogue of an integer $m$ is given by $[m]_q=(1-q^m)/(1-q)$. Let $a$ be an integer, and let $n$ be a positive odd integer. Via discrete Fourier transforms, we establish the following two identities: $$\det\left[\left[\left\lfloor\frac{aj-(a+1)k}n\right\rfloor\right]_q\right]_{1\leqslant j,k\leqslant n}=-\left(\frac{a(a+1)}n\right)q^{(1-3n)/2}$$ and $$\det\left[\left[\left\lceil\frac{(a+1)j-ak}n\right\rceil\right]_q\right]_{1\leqslant j,k\leqslant n}=\left(\frac{a(a+1)}n\right)q^{(n-1)/2},$$ where $(\frac{\cdot}n)$ denotes the Jacobi symbol.