All works

math.LO — Logic

The Logical Anatomy of Optimization: Reverse Mathematics, Proof Mining, and the Price of Convergence

Contributed by Xiaoyu Li

Optimization theory proves three kinds of theorems: that optimal points exist, that algorithms converge to them, and that optimality can be certified by duality. We measure how much mathematics each kind needs, using reverse mathematics, Weihrauch complexity and proof mining, and assemble the answers, old and new, into one map stated in the language of optimization. Compactness makes existence cost weak Kőnig's lemma, but convexity (and, format by format, tameness) makes it cheap: we show that $\mathsf{RCA}_0$ proves that convex functions on $[0,1]$ attain their minima (and on $[0,1]^n$, for each fixed $n$, given a modulus of uniform continuity), and that $\mathsf{RCA}_0$ proves linear programming duality and the minimax theorem for real data of every size. Convergence costs more: it follows from results of Kohlenbach and Neumann that convergence of gradient descent for convex smooth functions on $\mathbb{R}$ is equivalent to $\mathsf{ACA}_0$. We explain this by a selection principle: what costs is the particular minimizer the algorithm selects, and uniqueness or an error bound removes the cost. Below convergence lies finitary information. Gradient descent in $\mathbb{R}^d$ has at most $2C_dD/\varepsilon$ fluctuations of size $\varepsilon$, with $C_d$ depending only on $d$, which yields rates of metastability; no such bound is dimension-free, by the span-restriction argument behind performance estimation; and the limits of gradient descent on computable instances have d.c.e. coordinates, a proper subclass of the limit-computable reals. The map also covers Ekeland's variational principle (up to $\Pi^1_1\text{-}\mathsf{CA}_0$), separation theorems, Kőnig duality ($\mathsf{ATR}_0$), proof-mined rates for fixed-point iterations, and the feasible world, where local search captures the $\mathsf{NP}$ search problems provably total in $T^1_2$ and approximate stationary points of smooth functions are complete for $\mathsf{CLS}=\mathsf{PPAD}\cap\mathsf{PLS}$. We close with five principles and open problems.

math.NT — Number Theory

Almost all $k$-Markov numbers are unique

Contributed by Qiyue Tang

For every fixed nonnegative integer $k$, we prove that almost every positive integral solution of [x^2+y^2+z^2+k(xy+xz+yz)= 3(k+1)xyz]is uniquely determined, up to permutation, by its largest coordinate.

math.NT — Number Theory

Monotonicity of Erdős sums

Contributed by Jaideep Padhi

For $k \geq 1$ let $f_k = \sum_{\Omega(n)=k} 1/(n\log n)$, where $\Omega$ counts prime factors with multiplicity, and let $g_k$ be the same sum restricted to odd $n$. Lichtman showed that $f_k \to 1$, with a global minimum at $k=6$, and Gorodetsky, Lichtman and Wong that $f_k = 1 - 2^{-k}(a_1 k^2 + O(k \log(k+1)))$. We prove that $f_k < f_{k+1}$ for every $k \geq 6$, and that $g_k > g_{k+1}$ for every $k \geq 1$. The two statements come from one analytic distinction. Writing $\sum_k P_k(s) z^k = \zeta(s)^z H(s,z)$, at $s=1$ the point $z=2$ lies on the first singular circle of $H$, and strictly inside the first singular circle of the product with the prime $2$ removed. We show that this distinction decides the direction in general. If the complement of a set of primes $\mathcal{P}$ is thin and its least element is at least $3$, the corresponding sums eventually decrease, so the second theorem is one case of a general phenomenon. The inequalities used at the last step of the first theorem are certified by interval arithmetic.

math.QA — Quantum Algebra

Haagerup symmetry in a rank-six lattice conformal field theory

Contributed by Cédric ARHANCET

Let $U$ be a positive strongly rational vertex operator algebra of central charge eight with $\mathrm{Rep}(U) \cong Z(\cal H_3)$. We show that its Cartan Heisenberg commutant $\Omega$ embeds conformally into a lattice algebra $V_K$, where $K$ has rank six, determinant $39$ and root system $A_2\oplus A_2\oplus A_1$. The diagonal theory of $V_K$ has a unique vacuum and carries each of the three Haagerup fusion categories through three nonconjugate embeddings of $\Omega$. We show that, up to isomorphism fixing $\Omega$, there are exactly five conformal extensions of $\Omega$ with positive strongly rational target and global representation-category dimension $39$, and that their images in $V_K$ form four conjugacy classes. We also determine the common invariant relevant spectrum of the three canonical actions, their normalized cubic constants and their invariant genus-zero field algebra. We work in the TFT framework of rational full conformal field theory.

math.DS — Dynamical Systems

The necklace of a rational right triangle: exact counting and complete periodicity

Contributed by Joshua Boorstein

Let $T$ be a right triangle with acute angle $\alpha$, $2\alpha/\pi = P/Q$ in lowest terms. The Katok--Zemlyakov unfolding of $T$, quotiented by its holonomy involution, is a necklace of $Q$ kites: an explicit genus-zero half-translation surface in the stratum $Q(P-2, Q-P-2, -1^Q)$, whose kites glue in steps of $P$ and whose interfaces have explicit transverse measures. We use it to count the cells into which the vertices of $T$ cut the beam perpendicular to each side. For a completely periodic direction class the separatrix diagram has exactly $Q$ edges, and sorting them by the involution gives the identity $Q = h + f + 2p$ ($h$ sides of $T$ in the class, $f$ extra cells of the perpendicular beams, $p$ involution-swapped edge pairs). When $p = 0$ it is a covering law: the perpendicular beams of a class cut $Q - 1$ or $Q + 1$ cells in total at odd $Q$, according to the class, and $Q$ at even $Q$; with no hypothesis the shortfall is even, non-negative, and zero only on a completely periodic class. Complete periodicity of a class — that almost every trajectory in it meets a side at a right angle — is equivalent to $p = 0$ and to the covering law (on two-beam classes off the Veech locus, one direction assumes an orphan branch, which companion papers prove at broad classes of centres); it fails at $8/15$. For the beam perpendicular to a leg the count $n(P/Q)$ is at most $Q - 1$ at every centre and is known exactly on the Veech locus $\min(P, Q-P) \leq 2$; off it we prove $n(3/Q) \leq 2\cdot\text{round}(Q/3) - 1$ for $Q \geq 5$, with equality exactly under complete periodicity, and $n(5/(5m+1)) = 5$ for every $m \geq 3$. Finally, every periodic billiard path of a rational right triangle in a mirror direction meets a side at a right angle.

Author comments: First version. Companion papers: "The orphan theorem" (hexagon:2611.00191); "No arrivals" and "Non-perpendicular periodic orbits in mirror directions of right triangles" in preparation. Code: https://doi.org/10.5281/zenodo.24037744

math.NT — Number Theory

Consecutive Divisors of Sums of Fibonacci Polynomials: a Luca–Marques Analogue in Finite-Field Polynomial Rings

Contributed by Scott Kominers

For every prime power $q$ with $\gcd(q,6)=1$, we construct infinitely many pairs of consecutive monic polynomials in $\mathbb{F}_q[T]$ whose product divides the Fibonacci-polynomial sum with upper index one less than their common norm. This gives a function-field counterpart of the divisibility condition in the Luca–Marques theorem on consecutive integral arithmetic means of Fibonacci numbers. A short extension of the construction gives infinitely many triples of consecutive monic polynomials with the same property.

Author comments: Dedicated to the memory of Florian Luca

math.NT — Number Theory

Making the coprimality convention explicit in a mean value theorem of Halupczok for the ternary Goldbach problem

Contributed by Steve Hubbard

Theorem 1.2 of Halupczok (2013) is a Bombieri-Vinogradov type mean value theorem for the ternary Goldbach problem with one prime in an arithmetic progression and two primes in short intervals. Her proof works with residue classes $a \bmod q$ coprime to the modulus, as stated on p. 344 of her paper, and this convention needs to be written explicitly into the statement because the printed product and the prime sum differ at $a = q = 3$, where the prime sum is $0$ for $X_1 > 3$ while the printed product is at least $9/16$ for odd $n$. We give two closely related explicit formulations, Formulation A and Formulation B, show that Formulation A implies Formulation B, and prove both assuming two cited inputs, Hypothesis H1 and Hypothesis H2, or alternatively Theorems 2.1 and 2.3 of her paper, which follow from these two inputs. Here Hypothesis H1 is a uniform reading of a printed remark of Perelli, Pintz and Salerno, supported by a reading of the proof, and Hypothesis H2 is assembled from displays of Perelli and Pintz and is not a numbered statement. The supplement of Theorem 1.2 with a maximum over $a$ is not covered. The later results of Halupczok's paper use only coprime residue classes and are unaffected. The main results of this note have been checked by computer.

math.RT — Representation Theory

Lusztig's sharp bound for Hecke algebras

Contributed by Qiyue Tang

We prove Lusztig's sharp boundedness conjecture for Hecke algebras of finite-rank Coxeter systems with positive integer-valued weight functions.

Author comments: v2(2026/11/11)：Version 1 passed the Rethlas-OSV review, which used GPT-6-Astra at high reasoning effort with four concurrent workers. The review reported no mathematical issues. By posting this manuscript on this platform, I do not seek to assert ownership of the results or claim priority. I welcome anyone interested in preparing a finished account or checking the arguments; the eventual scholarly credit should go to those who undertake that work.

math.FA — Functional Analysis

Improved scaling in the metric cotype inequality

Contributed by Manor Mendel

Let $2\le q<\infty$. It is proved that every Banach space with Rademacher cotype $q$ satisfies the metric cotype $q$ inequality with scaling parameter $m=O(n^{1/2+1/q})$. The improved bound on the scaling is achieved by reducing the metric cotype problem to a smoothing and approximation criterion on the continuous torus. For a single unconditional kernel, the criterion is controlled by its mean $\ell_\infty$ radius and its variation in the main diagonal direction. The construction fits the universal smoothing and approximation scheme of Giladi, Mendel, and Naor. The proofs were generated by GPT~6 Astra with the help of the author.

Author comments: 17 pages

math.HO — History and Overview

The 57 Grothendieckian Problems: Open Problems for Mathematics and Machine Reasoning from the Grothendieckian Programme

Contributed by Vincent Liu

We state fifty-seven open problems drawn from the seven works of the Grothendieckian Programme (34 volumes, 26,021 pages, released openly in October 2026 as first drafts that have not been peer-reviewed). Thirty-five come from the two works on the architecture of a reasoning system and twenty-two from the five works that each take a named conjecture or programme as research subject; no problem is one of those conjectures. Each problem is stated in the words of its source, located by work, chapter and page, given the status the source assigns it, and paired with a criterion that decides when it is solved, a note on existing work and the date of a literature check. The problems are of three kinds: twenty-five mathematical problems, nineteen machine problems with success criteria fixed in advance, and thirteen formalization problems for Lean 4 and Mathlib. Every quotation from the sources has been checked against the published PDFs, and each problem has been checked against the literature as of 28 October 2026; a machine-readable version accompanies the paper.

Author comments: 54 pages, 5 figures. Version 1.0 of 28 October 2026; the same text is deposited at Zenodo, DOI 10.5281/zenodo.23734427 (all versions: 10.5281/zenodo.23734426). This copy was typeset with pdfLaTeX for Hexagon: its text is identical to version 1.0 at Zenodo, whose sealed PDF was typeset with LuaLaTeX, and a few line and page breaks differ. Each problem has a public issue at https://github.com/AlienMuse/grothendieckian-problems, which also holds the machine-readable data, the literature-check records and the Lean statement of Problem 49. Corrections recorded since version 1.0, including literature missed by the check for Problem 18: https://github.com/AlienMuse/grothendieckian-problems/blob/main/ERRATA.md. Companion papers: The Grothendieckian Programme: An Overview (DOI 10.5281/zenodo.23734421) and The Rising-Sea Theses (DOI 10.5281/zenodo.23734423).

math.RT — Representation Theory

Pretorsion theories and image-closed subcategories

Contributed by Haruhisa Enomoto

A pretorsion theory generalizes a torsion pair by allowing its two components to have a nonzero intersection, with morphisms factoring through the intersection playing the role of zero morphisms. In the module category of a finite-dimensional algebra, the first and second components of pretorsion theories are known to be the quotient-closed and submodule-closed subcategories, respectively. We show that the intersections of the two components are precisely the image-closed subcategories, namely those containing the image of every morphism between their objects, and that the pretorsion theories form a complete lattice. For a linearly oriented quiver of type A, we prove that the numbers of pretorsion theories and of image-closed subcategories are the Genocchi and median Genocchi numbers, respectively. Counting pretorsion theories by the number of indecomposable modules in their intersection, and image-closed subcategories by their number of indecomposable modules, yields the Gandhi polynomial refinements of Han and Zeng.

Author comments: Currently we are working Lean formalization, which will be released soon.

math.RA — Rings and Algebras

The L’vov–Kaplansky Conjecture for Matrices of Sufficiently Large Order

Contributed by Qiyue Tang

For each positive degree, the L’vov–Kaplansky conjecture holds over arbitrary fields in all matrix sizes beyond a threshold depending only on the degree. In these sizes the image of a nonzero multilinear polynomial is the traceless subspace if every cyclic coefficient sum vanishes, and the full matrix algebra otherwise.

math.RT — Representation Theory

Lusztig's conjectures for $B_n$, $\widetilde B_n$, and $\widetilde C_n$

Contributed by Qiyue Tang

We prove Lusztig's conjectures $\mathrm{P1}$–$\mathrm{P15}$ for every positive integer-valued weight function on the finite Coxeter group of type $B_n$ and the nonextended affine Weyl groups of types $\widetilde B_n$ and $\widetilde C_n$.

Author comments: v2:（2026/11/12）：Version 1 underwent a Rethlas-OSV review, including a supplementary round that completed coverage of all 36 review scopes. The final report recommended revision and classified all four findings as minor. It gave a counterexample to the assertion that the largest surviving cuspidal odd part always equals \(4u-2\). The report recommended using \(4u-2\) as an upper bound and establishing equality at the later step where it is needed. The other three findings concerned the attribution of the general parameter range in the introduction, the citation supporting the sharp asymptotic threshold, and the explanation of the parameter normalization used to obtain the residual centers. The report proposed a specific correction for each finding and identified no issue requiring a change to the main conclusions. The revised manuscript underwent a second Rethlas-OSV review, which identified only two issues: an incorrect citation in the introduction and an inaccurate article title in the bibliography. I am posting this manuscript to share the mathematical results themselves. In doing so, I do not seek to assert ownership of the results or claim priority. I welcome anyone interested in checking the arguments or preparing a finished account, and I believe the eventual scholarly credit should go to those who undertake that work.

math.NT — Number Theory

Odd/Odd Records for a Relative Approximation Error: Three Quadratic Classifications and an Exact Certificate

Contributed by Elias De Jesús

For a quadratic irrational α > 1 and positive odd integers m, k, set H = mα + k and ε = |mα − k|/(mα + k). A state (m, k) is a strict chronological record if ε is smaller than at every admissible state of smaller H. We classify the strict records completely for α = φ, α = 1 + √3 and α = √3. In each case they are a finite explicit union of orbits of a single unimodular matrix, with signed defect cycles (−1, +1, −5), (−3, +1, −11) and (−2) respectively. The proofs rest on two exact identities that make ε a function of the defect N and the conjugate factor η alone, on a norm cutoff |N| ≤ 18, 204, 27 derived from heights, and on a finite exact certificate that decides every remaining comparison. Two phenomena require care and are isolated: at φ the competition between the Fibonacci and Lucas ladders is a genuine tie at leading order, decided by an integer identity N_B = −5N_A; and at √3 a companion family has exactly equal error at larger height, so it is excluded by equality rather than by loss. The results resolve two conjectures of the author's earlier note and correct one assertion there. Nothing here is formally verified or refereed, and the uniform generalisation is left open.

Author comments: First version, 17 pages. Supporting materials—including the exact-arithmetic certificate, verification scripts, complete orbit tables, and supplementary documentation—are available at: https://github.com/innerlightr-wq/two-mode-oscillator-record-chronologies/tree/main/congruence-records

math.NT — Number Theory

Subrings of $\mathbb{Z}^n$ of prime-power index and modular forms

Contributed by Anwesh Ray

Let $f_n(m)$ denote the number of unital subrings of index $m$ in $\mathbb{Z}^n$, with coordinatewise multiplication, and let $g_n(p^e)$ count those whose reduction modulo $p$ is the diagonal copy of $\mathbb{F}_p$. Voll's uniformity question asks whether the local subring zeta functions can be described by a single rational function in $p$ and $p^{-s}$. Polynomial formulas for $f_n(p^e)$ are known for $e\le9$. We prove that neither $f_{10}(p^{17})$ nor $g_{10}(p^{17})$ agrees with a polynomial in $p$ for all sufficiently large primes. We introduce new geometric and cohomological methods for this problem, moving beyond the earlier combinatorial calculations. We compare the counts over $\mathbb{Z}_p$ and $\mathbb{F}_p[[t]]$ and, for all sufficiently large primes, express them as point counts on reductions of varieties over $\mathbb{Q}$. The Grothendieck--Lefschetz trace formula relates these counts to compactly supported $\ell$-adic \'etale cohomology. We identify a contribution from Tate twists of the two-dimensional Galois representation attached to the weight-twelve cusp form $\Delta$. We show that this modular contribution does not cancel when all parts of the count are combined. Consequently, there is no polynomial $P\in\mathbb{Q}[X]$ such that $f_{10}(p^{17})=P(p)$ for all but finitely many primes $p$. This gives a negative answer to Voll's uniformity question even when finitely many exceptional primes are allowed.

math.CO — Combinatorics

The Kopotun Subpermanent Transform Inequality

Contributed by Yair Lavi

We prove the subpermanent transform inequality conjectured by Kopotun in 1996: for every doubly stochastic matrix $A$ of order $n\ge2$ and every $2\le k\le n$, $\sigma_k(A)\ge\sigma_k((nJ_n+A)/(n+1))$. Here $\sigma_k$ is the sum of all subpermanents of order $k$, and $J_n$ has every entry $1/n$. Equality holds exactly at $A=J_n$.

cs.LO — Logic in Computer Science

Formalizing the interpolation lemma of the openai/math Landau–Siegel paper and substituting it into a Comparator-accepted proof

Contributed by Joseph Reilly

OpenAI’s openai/math release includes a Lean formalization of “Uniform exclusion of Landau–Siegel zeros” (OpenAI, October 1, 2026). Its Comparator challenge has two checked theorems, whose statements we prove equivalent in Lean. OpenAI proves them by two routes, neither through the paper’s proof of Lemma 3. The catalogued main result, exists_absolute_real_zero_gap, gets its spanning step from a local Bézout argument and does not use OpenAI’s Corollary 4 theorem. The second theorem, dirichletRealZeroBound_proof, proves Corollary 4 by a multiplicity estimate. We formalized the paper’s proofs of Lemma 3 and Corollary 4 in Lean. By reading, the Lean statements match the paper’s, and the declarations use only the axioms propext, Classical.choice and Quot.sound. A +2/-1 substitution puts the paper’s Corollary 4 into dirichletRealZeroBound_proof. Comparator accepts the result with both kernels at openai/math fd4aeeb under a deny list of all 45 socket address families, and that theorem’s closure no longer contains the three named OpenAI lemmas. Our replay of the corpus agrees with an independent replay on all 402 configurations where both have a verdict.

math.CO — Combinatorics

An affine criterion for a classical four-point family

Contributed by Alexandre Couret

For a classical family of four-point subsets of a cyclic group, we give an explicit greatest-common-divisor criterion for equivalence under translations and arbitrary unit multipliers. Let $q>0$ and $x\in\mathbb{Z}/(4q)\mathbb{Z}\setminus\{0,q,2q,3q\}$, with representative $X$. The supports $P=\{0,q,x,x+2q\}$ and $Q=\{0,q,x+q,x+3q\}$ have identical labelled ordered sum profiles, including diagonal pairs. They are not affinely equivalent exactly when $q/\gcd(q,X)\equiv2\pmod4$. The argument handles nonunit parameters by lifting units between quotient rings. No bibliographic novelty is claimed.

Author comments: Short note for mathematical and bibliographic review. The four-point family is classical; no novelty claim is made for the affine criterion. Independent human mathematical review remains outstanding. Ancillary Python code provides a finite check for q <= 32 and does not replace the general proof.

math.CO — Combinatorics

Complex and real counterexamples to an inequality arising from Drury's conjecture

Contributed by Yair Lavi

We give counterexamples to the following inequality, stated by Zhang as a consequence of Drury's 2016 permanent conjecture: if $A=(a_{jk})_{j,k=1}^n$ is a Hermitian positive semidefinite matrix of order $n\geq2$, then $\sum_{j=1}^{n-1}|a_{jn}|^2\operatorname{per} A[\,\hat{\jmath},\hat{n}\,]\leq\operatorname{per} A$ where $A[\,\hat{\jmath},\hat{n}\,]$ is the principal submatrix obtained by deleting indices $j$ and $n$. Our complex example has order twelve and rank two; our real example has order twenty-one and rank three.

math.CO — Combinatorics

A Dittert-like inequality for homogeneous stable polynomials

Contributed by Yair Lavi

We prove a coefficient inequality for every homogeneous real stable polynomial $P(s,x_1,\ldots,x_n)$ of degree $n\ge1$ with nonnegative coefficients. If $P(1,0,\ldots,0)=1$, $r_i=\partial_{x_i}P(1,0,\ldots,0)$, and $\sum_i r_i=n$, the following inequality holds, where brackets denote coefficient extraction: $[x_1\cdots x_n]P(0,x_1,\ldots,x_n)+2-\prod_i r_i-P(0,1,\ldots,1)\ge \frac{n!}{n^n}$. Equality holds exactly for $P=(s+\sum_i x_i/n)^n$. The normalizations can be removed by rescaling $P$ and the $x_i$. We give applications to permanents, mixed discriminants, subpermanents, and stable combinatorial measures. In particular, the theorem gives an independent proof of Dittert's conjecture for nonnegative matrices in every dimension.

cs.DS — Data Structures and Algorithms

An Optimal Deterministic Adversary for Sorting

Contributed by Nicholas Nash

This paper describes an optimal deterministic, polynomial-time adversary for ordinary sorting. For any correct sorting algorithm, the adversary forces that algorithm to make $(1-o(1))n\log_2 n$ comparisons. The best previously known deterministic polynomial-time adversary forced $\frac{37}{64}n\log_2 n$ comparisons. The adversary maintains a family of orders consistent with the answers so far, using a hierarchy of directed acyclic graphs. Bounds on vertex degrees and directed path lengths allow comparison answers to be chosen through deterministic calculations on small graph neighborhoods.

Author comments: Substantially revised and expanded exposition, with explicit algorithms, intuitive explanations of the main proofs, and further consequences. 36 pages. Source and Lean formalization: https://github.com/nicknash/sorting-adversary-proofs .

math.RA — Rings and Algebras

The L’vov–Kaplansky conjecture in degrees three and four over arbitrary fields

Contributed by Qiyue Tang

We prove that for every field (K) and every positive integer (n), the image of each multilinear noncommutative polynomial of degree three or four on (M_n(K)) is a (K)-linear subspace.

math.RT — Representation Theory

The singular-support characterization of Lusztig sheaves

Contributed by Qiyue Tang

We prove that for every finite loop-free quiver and every dimension vector, a simple equivariant perverse sheaf with complex coefficients on the quiver representation space is a Lusztig sheaf if and only if its singular support lies in Lusztig’s nilpotent variety.

Author comments: v2（2026/11/11）：Version 1 passed the Rethlas-OSV review using GPT-6-Astra (high reasoning effort, four concurrent workers). The review reported one incorrect theorem number in a citation and no mathematical errors. By posting this manuscript on this platform, I do not seek to assert ownership of the results or claim priority. I welcome anyone interested in preparing a finished account or checking the arguments; the eventual scholarly credit should go to those who undertake that work.

math.CV — Complex Variables

An explicit counterexample of degree 29440 to Smale's mean value problem with constant one

Contributed by David Niedbala Giraudin

For a complex polynomial p with p(0)=0 and p'(0)=1, let S(p,0) be the minimum of |p(c)/c| over the critical points c of p. Adamczewski and Colbrook recently gave an explicit polynomial of degree 29984 with S(p,0)>1, certified by interval arithmetic, built on the counterexample found by GPT-6 Astra. Within their family of polynomials, with re-chosen coefficients, we give an explicit p in Q(i)[z] of degree 29440 with 1.00000093069254310 < S(p,0) < 1.00000093069254311. Jatar and Ng have shown that Smale's constant sup_p S(p,0) exceeds 1.1018, through polynomials of unspecified degree; the point here is an explicit polynomial with a certified bound. The bound is certified twice: by a ball-arithmetic Krawczyk certificate covering all 29439 critical points, and, independently, by the Adamczewski-Colbrook verifier, which confirms 29418 of them and is inconclusive (insufficient working precision) on the other 21. The exact coefficients and both verification programs are provided as ancillary files.

Author comments: 4 pages. Ancillary files: exact specification (polynomial_N1840.json), critical-point centres and disks, ball-arithmetic verifier (verify_ball_N1840.py), certificate summary, and the parameter patch used to run the Adamczewski-Colbrook verifier.

math.RT — Representation Theory

Self-injective dimensions under stable equivalences of Morita type

Contributed by Qiyue Tang

Stable equivalences of Morita type preserve the left and right self-injective dimensions of Artin algebras. The defining tensor functors can send injective modules to noninjective modules.

math.AC — Commutative Algebra

Counterexamples in positive characteristics

Contributed by Rahul Ajit, Seungsu Lee

AI found the following counterexamples (following human directions) to 3 important questions: For every prime $p$ we construct a smooth projective rational surface that is globally $+$-regular, with anticanonical divisor nef but neither big nor semiample. Its completed cone is a three-dimensional $F$-finite splinter with regular punctured spectrum and embedding dimension six. Both test ideals are the maximal ideal, so the cone is not weakly $F$-regular. A discrepancy calculation by humans rules out every effective klt boundary. A separate four-dimensional complete normal local domain has distinct big and finitistic test ideals.

Author comments: All main ideas and constructions in this draft are either explicitly suggested or fully developed by AI. The humans (R. Ajit and S. Lee) asked specific questions and gave related directions, papers, and even precise result statements to AI so that it could explore those directions and construct counter-examples. In most cases, AIs (and humans) were successful. However, their mathematical details were quite incomprehensible to the humans. Currently, Ajit and Lee are carefully verifying AI's claims as well as improving the exposition and adding their personal human inputs. What follows is a very preliminary attempt to that. In particular, the humans think the math in this draft is correct. In the near future, with a better understanding, the humans think that they will be able to substantially improve the exposition, add and correct proper citations, and most importantly add precise human contributions (right definitions, for example) to this project. We will also add more precise human contributions to the existing material soon. We make this available to the world purely for mathematical reasons.

math.DS — Dynamical Systems

The orphan theorem: degenerate perpendicular orbits in rational right triangles

Contributed by Joshua Boorstein

Let $T$ be a right triangle with acute angle $\alpha$, legs $L_1$, $L_2$ and hypotenuse $H$, and suppose $T$ is rational: $2\alpha/\pi = P/Q$ in lowest terms. Shoot a beam perpendicular to the leg $L_1$ from each of its points. The leg is partitioned into branches — maximal intervals of launch heights sharing a combinatorial type — and we prove that the partition is finite, with at most $2Q + 1$ branches, which cover the leg but for finitely many points. The first-return map $J$ of the beam to $L_1$ is an involution, and on each branch it does one of two things: it carries the branch onto a distinct branch of equal width by a reflection, or it fixes the branch pointwise. A pointwise-fixed branch — a whole interval of trajectories that hit a side head-on and retrace — is an orphan. Theorem A. There is no orphan branch when $P$ is even and at most one when $P$ is odd, with no hypothesis on the direction; when $P$ is odd and $Q$ is a prime power, or twice or four times a prime, the orphan exists. Non-existence is a parity count on the velocity set of the unfolded beam. Uniqueness is an Euler count on the genus-zero quotient of the unfolded surface, which admits minimal components. Existence is arithmetic: on a folded disc, an absent orphan forces $1$ to be a signed sum of doubled heights of images of the right-angle vertex, which is impossible at the centres named. Running the direction count and the Euler count for each side shows that at most two of the three sides carry a retracing cell, at most one each, the two ends of one band of double normals. On the unfolded surface — the holonomy double cover of a genus-zero quadratic differential $Q(P-2, Q-P-2, -1^Q)$ — the images of the right-angle vertex are Weierstrass points, the leg is the quotient by the hyperelliptic involution of a geodesic segment through one of them, and the first-return map is read in that quotient.

Author comments: First version. Companion papers in preparation: "The necklace of a rational right triangle" and "No arrivals". Code: https://doi.org/10.5281/zenodo.24037744

math.QA — Quantum Algebra

Positivity of Canonical Basis Multiplication in Modified Quantum Groups

Contributed by Qiyue Tang

We prove the multiplication part of Lusztig’s positivity conjecture for symmetric Cartan data.

Author comments: v2（2026/11/11）：Version 1 passed the Rethlas-OSV review, which used GPT-6-Astra with high reasoning effort and six concurrent workers. The review identified a single bibliographic error. By posting this manuscript on this platform, I do not seek to assert ownership of the results or claim priority. I welcome anyone interested in preparing a finished account or checking the arguments; the eventual scholarly credit should go to those who undertake that work.

math.AG — Algebraic Geometry

The tangent line at a marked point on $M_{0,n}/\mathfrak S_{n-1}$, after a table of Grothendieck

Contributed by Michel Hua

Let $L_n$ be the line bundle on $M_{0,n}$ whose fibre is the tangent line to $\mathbb P^1$ at the last marked point, and $L$ the tensor product of the tangent lines at all of them. We show that, in characteristic zero, the Picard group of the quotient stack $[M_{0,n}/\mathfrak S_{n-1}]$ is cyclic of order $(n-1)(n-2)$ for every $n \ge 3$, generated by $L_n$, and that $L$ has order $n-1$ in $\operatorname{Pic}[M_{0,n}/\mathfrak S_n]$. For even $n \ge 6$ the group was known, through the integral Chow ring computed by Edidin and Hu; the odd case does not seem to be in print. The question comes from a table in Grothendieck’s working notes, which gives these orders for four points and only guesses for five and six. The proof takes a few lines once the stack is written as monic squarefree polynomials modulo the affine group. An elementary version, with explicit trivialisations and fixed configurations, is formalised in Lean.

math.CO — Combinatorics

A bigon-free dichotomy for pseudoline arrangements, after a note of Grothendieck

Contributed by Michel Hua

A pseudoline of an arrangement in the real projective plane is bigonal when it carries exactly two vertices of the arrangement. In an undated folder of working notes from the early 1980s, Grothendieck states that every arrangement without bigonal pseudolines is either four pseudolines in general position or contains one of three configurations: five pseudolines in general position, a pencil of three and a pencil of two without the line joining their centres, or the complete quadrangle. The notes derive this from a statement that is false as written, and they give no independent proof. We give a short proof, together with the ordering of the pseudolines that the notes ask for, and we deduce that for n ≥ 6 the near-pencils are the only arrangements of n pseudolines containing none of the three configurations. The dichotomy has been checked by an exhaustive enumeration of all arrangements of at most nine pseudolines, and it is formalised in the proof assistant Lean for a combinatorial model of arrangements, generalised maps of Euler characteristic one.

math.CO — Combinatorics

Counterexamples to Source–Sink Recursions for Denominator Vectors

Contributed by Qiyue Tang

Reading and Stella’s source–sink recursions for denominator vectors, Conjectures 2.7 and 2.8, fail for an infinite family of four-vertex exchange matrices. The member realized by an unpunctured annulus with one marked point on one boundary and three on the other also refutes their Theorem 2.11.

math.RT — Representation Theory

Singleton Dominance Regions in Affine Cluster Algebras

Contributed by Qiyue Tang

For an extended exchange matrix of affine type, the dominance region of every point in an extended (g)-vector cone consists of that point alone.

math.RA — Rings and Algebras

The L’vov–Kaplansky Conjecture for Standard Polynomials

Contributed by Y. T. Miah

The L'vov–Kaplansky conjecture asks whether the image of a multilinear polynomial on a full matrix algebra is a vector subspace. For the unnormalized standard polynomial $s_d$ over an algebraically closed field $K$ of characteristic zero, the sets of individual matrix values are determined in every degree and order. The image of $s_{2r-1}$ is zero when $n<r$ and all of $M_n(K)$ when $n\ge r$; the image of $s_{2r}$ is zero when $n\le r$ and $\mathfrak{sl}_n(K)$ when $n\ge r+1$. The odd-degree proof uses explicit linear slices and Jordan forms. The even-degree argument combines a uniform nilpotent construction and an additive transfer for admissible cores with generic-nullity results of Brassil–Reichstein and Lin–Tang–Zhao, extending a method of Qiyue Tang to obtain actual polynomial values.

math.DS — Dynamical Systems

Globally periodic rational difference equations of arbitrary order

Contributed by John Rudge

A rational difference equation of order $k$ with nonnegative coefficients is globally periodic if every solution with positive initial data repeats with a common period. Five such equations have long been known, among them the Lyness equation $x_{n+2}=(1+x_{n+1})/x_n$, and a problem of Gasull asks whether every globally periodic equation of this kind is obtained from one of the five by rescaling and unfolding. The answer was known to be yes for $k\in\{1,2,3,4,5,7,9,11\}$. Here I show that it is yes at every order. The proof rests on two ideas. The first is a tropical degeneration: global periodicity forces an associated min-plus recursion to be periodic, and this is refuted, for every choice of coefficients at once, by an unbounded solution of the min-plus recursion linearised along a periodic orbit. The second is the $\sigma$-bump, a linearised solution whose existence reduces to a combinatorial condition on residues modulo $2k$; in the hardest case that condition comes down to a lemma on three-term progressions under unit multipliers modulo $2k$. The proof, including its finite computations, has been formalised in the Lean proof assistant.

math.CO — Combinatorics

Height-bounded mutation to acyclic quivers on four vertices

Contributed by Qiyue Tang

We prove Conjecture 5.5 of Armstrong-Williams, Hirst, Jackson and Lee.

math.CO — Combinatorics

Complexity of Mutation-Acyclicity and Matrix Mutation Equivalence

Contributed by Qiyue Tang

Mutation-acyclicity of integral skew-symmetrizable matrices belongs to NP. Mutation equivalence belongs to promise-NP when at least one input is mutation-acyclic. For pairs of mutation-cyclic matrices, mutation equivalence is RE-complete and strongly NP-hard. Both results hold even for connected skew-symmetric matrices.

math.RA — Rings and Algebras

The Lu–Wenzel Conjecture

Contributed by Qiyue Tang

We prove the Lu–Wenzel conjecture for real and complex matrices.

cs.DS — Data Structures and Algorithms

A certified exponent of 0.48758 for worst-case Subset Sum, and limits of the alias–checksum analysis

Contributed by Jennifer Voitle

For more than fifty years, since Horowitz and Sahni's meet-in-the-middle algorithm of 1974, the best exponent for worst-case Subset Sum was $1/2$. In October 2026 that barrier fell twice within a week, with OpenAI's $2^{0.49n}$ algorithm and, independently, Li and Ye's $2^{(1/2 - 1/450000)n} \approx 2^{0.4999978n}$, and improvements have followed quickly since: within three weeks A. Kapila's modification of the OpenAI approach ("Base") reached 0.4877, a retuning of Base by JoshuaQFEX reached 0.487676, and a refinement by Y. Bills reached 0.4876723. Here we push this line of work further and examine what limits it. Conditional on the cited non-numerical proofs, we give a certified parameter point with threshold $T=0.4875754036$: for every fixed $T'>T$, the uniform randomized algorithm uses $O(2^{T'n})$ word operations on inputs of polynomial bit length, with success probability at least $2/3$. In particular, it runs in time $O(2^{0.4875755n})$. A single asymmetric filter replaces Base's family of 43 filters. Independent interval computations certify all required numerical conditions, including both filtering moments and their quadrature errors. We prove a lower bound on the first filtering moment, valid for all positive continuous densities, that excludes every parameter choice with $\tau\le0.484$ in the original OpenAI proof's sufficient budget. Three algebraic bounds explain the costs at our improved point. Separately, an enumeration phase that inserts at least $2^{(1-\eta-o(1))n}$ alias classes, spending a counted step on each, and uses checksum gap at most $\Delta$ needs exponent at least $(1-\Delta)/2$. This is a conditional operation-count bound; a global upper bound on achievable checksum capacity remains open. Finally, we isolate a small-layer statement that would yield threshold $0.4873407528$, prove combinatorial special cases, and identify the remaining questions.

Author comments: 30 pages, 2 figures. The main result is conditional on the written proofs of OpenAI (4 Oct 2026) and A. Kapila (commit 44c34961); its numerical conditions are certified in interval arithmetic. The companion file SECTION_T_v3.pdf (full proofs of the ternary alias step, Appendix A) is included with the source. Code, certificates, verification scripts and a timestamped research record: https://github.com/jv1000/subset-sum. AI use is described in Section 11.

math.CO — Combinatorics

Four-cycle supersaturation and polarity stability at $q^2+q$ vertices

Contributed by Zhenyu Sun

Let $m_0=\frac12q(q+1)^2-q$. For every even $q\ge1600$, we prove that a graph on $q^2+q$ vertices with at least $m_0+1$ edges contains at least $\lceil(q-1)/12\rceil$ four-cycles. For powers of two in this range, $m_0$ is the extremal number and a polarity construction has exactly $q-1$ four-cycles at $m_0+1$ edges. These bounds determine the order of magnitude of the minimum count on this sequence. We also prove a stability theorem: for every nonnegative integer $b$ and every even $q\ge10^{18}(b+1)^{16}$, each $C_4$-free graph on $q^2+q$ vertices with $m_0-b$ edges is a spanning subgraph of an orthogonal polarity graph with one point deleted. The required degree bounds follow from the edge count. As a consequence, a graph on $q^2+q$ vertices with $m_0+1$ edges has at least $q-1$ four-cycles if an edge set of size $t\ge1$ meets every four-cycle and $q$ is even with $q\ge10^{18}t^{16}$. The counting proof uses a codegree commutator; the stability proof combines degree-deficit concentration with incidence completion and polarity extension.

Author comments: 64 pages.

math-ph — Mathematical Physics

The Hawking-Page transition at weak coupling

Contributed by Anthropic

In anti-de Sitter space a black hole forms through a first-order phase transition, discovered by Hawking and Page. In $\mathcal{N}=4$ super Yang-Mills theory on $S^3$ this is deconfinement at strong coupling. Does the transition remain first order at weak coupling, where there is no classical geometry? The free theory has a transition at its Hagedorn temperature, and it was suggested in 2003 that the two could be continuously connected, either by a single first-order line or through a pair of continuous transitions at weak coupling, which have no counterpart in classical gravity. One number decides: the sign of the quartic coefficient $b$ in the effective action of the Polyakov loop. It first appears at three loops and, to our knowledge, has not been computed in the two decades since. We find $b=-1.48\times10^{-3}\,\lambda^2$, where $\lambda$ is the 't Hooft coupling. The transition is first order. It lies slightly below the Hagedorn temperature, and its three saddle points appear, exchange dominance and disappear in the same order as thermal anti-de Sitter space and the small and large black holes. The structure of the Hawking-Page transition is thus already present in perturbation theory. The computation is Hamiltonian, and almost every number in it was obtained in two independent ways. It reproduces the two-loop Hagedorn temperature known from integrability and the exact results available in closed subsectors.

math.RT — Representation Theory

Disproof of Xi’s conjecture on the representation dimension of idempotent corners

Contributed by Qiyue Tang

We disprove Xi’s conjecture on the representation dimension of idempotent corners. Over every field and for each integer (L\ge3), we construct an algebra of representation dimension (L+1) with a corner of representation dimension (2L-1).

math-ph — Mathematical Physics

The spinless modular bootstrap at large central charge

Contributed by Anthropic

The spinless modular bootstrap bounds the dimension $\Delta_1$ of the lightest primary other than the vacuum in a two-dimensional conformal field theory, using only modular $S$ invariance of the torus partition function at zero angular potential. For unitary, compact theories with central charge $c>1$ and no conserved currents beyond the Virasoro algebra, these bounds read $\Delta_1\le c/\Lambda+o(c)$ at large $c$. The best slope $\Lambda$ was known only to lie between $8.503$ and $12$, where the bound would reach the BTZ black hole threshold. We show that it is exactly $\Lambda_V=12\pi^2/(\pi^2+4\,\mathrm{arcsinh}^2 1)\approx9.1266$. No linear functional in a class that contains all finite combinations of derivatives improves on $c/\Lambda_V$ by more than $O(c^{1/3}\log^{2/3}c)$, and an explicit functional gives $\Delta_1<c/\Lambda_V+0.483\,c^{1/3}+O(1)$ for all sufficiently large $c$. The value has an electrostatic origin: as a function of a complexified dimension, $c^{-1}\log\lvert g\rvert$ of a functional $g$ is a two-dimensional potential with positive charges at the zeros of $g$. Positivity forces $g$ to fall as fast as the Cardy density of states grows, and the maximum principle bounds the potential by an explicit harmonic function, which fixes $\Lambda_V$. The field of this potential is the spectral curve of a mean-field theory of the optimal functionals, and its charge density is the distribution that the zeros of any functionals attaining $\Lambda_V$ must follow. We also conjecture the coefficient of the $c^{1/3}$ correction to the optimal bound, supported by numerics up to $c=2048$.

math.RA — Rings and Algebras

The L'vov--Kaplansky conjecture for multilinear polynomials of degrees four to six

Contributed by Shengyong Pan

Let $K$ be an algebraically closed field of characteristic zero. We determine the images of ordinary multilinear polynomials of degrees four, five and six on matrix algebras. For $n\geq3$, a nonidentity has image $\mathfrak{sl}_n(K)$ when its cyclic class is zero and $M_n(K)$ otherwise. In these degrees the only nonzero identity exception is the standard degree-six line on $M_3(K)$. The proof starts with proper sextics and treats arbitrary sextics by unit restrictions and local trace lifting. An image-preserving product substitution then gives degrees four and five as consequences. We also give cyclic and matrix lifting results applicable in arbitrary degree. The finite calculations are given by explicit matrix-unit identities.

cs.FL — Formal Languages and Automata Theory

Degree Transport and a Smaller Cubic Coefficient for Reset Thresholds

Contributed by SUNGBIN HONG

We prove that every synchronizing complete deterministic finite automaton on $n\ge2$ states has reset threshold at most $\frac{8242}{49923}n^3+O(n^2)$. The cubic coefficient improves on Shitov's $8257/49923$ by $5/16641$. The main ingredient is a degree-transport inequality for graphs of state pairs mergeable by short words. It relates a long gap between consecutive minimum-rank word lengths to a large corank achievable in at most $n-1$ letters. Combining this relation with a discrete dual of Shitov's counting argument gives the improvement. An appendix makes the quadratic remainder explicit for every $n\ge2$. The argument applies to arbitrary finite alphabets and does not prove the quadratic Černý conjecture.

math.NT — Number Theory

Four exponentials at CM and Beta periods

Contributed by Shawn Krisman

For every nonzero eigen-period $P$ of a CM abelian variety over $\overline{\mathbb{Q}}$, we show that the four exponentials conjecture and its strong form hold at $t\in\overline{\mathbb{Q}}^\times P^{\pm1}$ and $t\in\overline{\mathbb{Q}}^\times(P/\pi)^{\pm1}$; these include Beta values such as $t=\Gamma(1/5)^2/\Gamma(2/5)$. For an elliptic curve $E$ over $\overline{\mathbb{Q}}$ with complex multiplication, we also show that $e^{\pi^2}$ and $\wp_E(1/\pi)$ are not both algebraic. Further results concern two CM elliptic curves: for example, $\pi,\varpi,\Omega_\rho,\varpi\Omega_\rho$ are linearly independent over $\overline{\mathbb{Q}}$, where $\varpi$ and $\Omega_\rho$ are periods of $y^2=4x^3-4x$ and $y^2=4x^3-4$. These are explicit consequences of published theorems of Waldschmidt (1988) and Roy–Waldschmidt (1993), together with classical results of Baker, Schneider, Masser and Wüstholz; the contribution is the statements and the proofs written out at these periods. Cases at specific periods, such as that $2^{\varpi}$ and $3^{\varpi}$ are not both algebraic, are checked in Lean 4 and reported separately.

math.NT — Number Theory

A Lean formalization of Waldschmidt's subgroup theorem with period term for groups with elliptic factors

Contributed by Shawn Krisman

We formalize in Lean 4 with Mathlib Waldschmidt's algebraic subgroup theorem with period term (Theorem 4.1 of his 1988 Durham paper) for $G=\mathbb{G}_{\mathrm{a}}^{d_0}\times\mathbb{G}_{\mathrm{m}}^{d_1}\times E_1\times\dots\times E_m$, with $E_j$ elliptic curves with algebraic invariants, under the additional hypothesis $2N\ge d_1+2d_0+\kappa$. The proof uses only Lean's three standard axioms. Along the way we give what we believe are the first formal proofs of Schneider's theorems on elliptic functions and of the Schneider--Lang criterion, and the first Lean library for the Weierstrass $\sigma$ and $\zeta$ functions. Restricted forms of the theorem cover some vector extensions and the Jacobian of $y^5=x(1-x)$. A companion certificate uses these to prove explicit cases of the four exponentials conjecture: for example, $2^{\varpi}$ and $3^{\varpi}$ are not both algebraic, where $\varpi$ is the real period of $y^2=4x^3-4x$, and likewise $2^{B(1/5,1/5)}$ and $3^{B(1/5,1/5)}$.

math.CO — Combinatorics

Foregger's Conjecture on Minimum Permanents of Nearly Decomposable Doubly Stochastic Matrices

Contributed by Yair Lavi

We prove Foregger's conjecture: every nearly decomposable doubly stochastic matrix $A$ of order $n\ge2$ satisfies $\operatorname{per}A\ge 2^{1-n}$. Equality holds exactly for matrices obtained from $\tfrac12(I_n+P_n)$ by independent row and column permutations, where $P_n$ is a single-cycle permutation matrix. In fact, for every fixed nearly decomposable support whose bipartite graph is not a single cycle, the minimum permanent on its closed doubly stochastic face is strictly larger than $2^{1-n}$.

math.DG — Differential Geometry

HOLOMORPHIC SECTIONAL CURVATURE OF FLAG MANIFOLDS VIA KÄHLER REDUCTION

Contributed by Ching Hsien Lee

We give an alternative proof that every invariant Kähler metric on a compact generalized flag manifold has positive holomorphic sectional curvature. Invariant Kähler metrics on unitary partial flag manifolds are realized as reductions of Hermitian vector spaces at compact moment-map levels contained in spheres. A curvature identity for linear Kähler quotients gives an explicit positive lower bound. For an inclusion of compact connected Lie groups, we prove that the induced inclusion between adjoint orbits through the same Lie algebra element is holomorphic, isometric, and totally geodesic for the Kähler structures determined by an invariant inner product and its restriction. This reduces the general case to unitary flags. We also obtain a strict estimate for the second fundamental form of the anticanonical projective embedding

math.CO — Combinatorics

On a counting conjecture for binary rich words

Contributed by Max Glass

Let $R_2(n)$ denote the number of binary rich words of length $n$. We prove that $\limsup_{n\to\infty}\frac{\log R_2(n)}{\sqrt{n}\log n}=\infty$. Consequently, $R_2(n)$ is not $O(n^{C\sqrt{n}})$ for any fixed $C>0$, disproving the conjectured upper bound $O((n/g(n))^{\sqrt{n}})$ with $g(n)\to\infty$. The proof combines independent permutation families through an ordered concatenation construction and a binary gap encoding.

Author comments: Formatting correction to the web abstract: repaired math delimiters and removed hard line breaks. The paper and mathematical results are unchanged. 4 pages. Lean 4 formalization and paper source: https://github.com/glasss13/rich-word-counting.

math.CO — Combinatorics

The constant Steiner kernel is not optimal at large girth

Contributed by Matthew Kwan, Huy Tuan Pham

In the paper "nonlinear lower-tail large deviations at criticality", we bound the number of order-$n$ Steiner triple systems with no Pasch configuration. In doing so, we show that a certain variational problem is minimised by a constant kernel. Using ChatGPT 6 Pro, we also discovered that the analogous variational problems for Steiner triple systems of girth at least $g$, for large $g$, are *not* minimised by the constant kernel, and we included a brief remark to this effect. In this note (not intended for publication), we present the details of this proof.