math.RT — Representation Theory
Broué's Abelian Defect Group Conjecture for 2-Blocks
We prove that every 2-block of a finite group with an abelian defect group is splendidly Rickard equivalent to its Brauer correspondent.
Browse
Browse the Hexagon catalog.
math.RT — Representation Theory
We prove that every 2-block of a finite group with an abelian defect group is splendidly Rickard equivalent to its Brauer correspondent.
math.CT — Category Theory
Let and let be a bounded coherent -topos. Suppose that finitely many noninitial coherent -truncated objects cover its terminal object, and that every coherent -truncated object in each of their slices has a finite covering family of sections. We identify with for a -finite -type , recovered by all finite models preserving finite limits, finite coproducts, and effective epimorphisms. Descent uses degrees through for objects and for maps; both bounds are sharp for finite homotopy actions. Every exact functor between recognized categories has a universal parameter and a canonical factorization through its slice with fully faithful second functor. This parameter characterizes full faithfulness, conservativity, and all exact slice factorizations. For successive abelian Postnikov layers, we identify the space of marked exact lifts with a path space of Eilenberg--Mac Lane sections with local coefficients. The obstruction, classes of lifts, and relative automorphisms occur in degrees , , and . We compute a twisted parity obstruction and, for odd primes, classify cyclic self-equivalence components by a -adic valuation. A nonzero Postnikov class can persist in the identity component. For continuous finite families over a profinite -type, we prove that a uniform finite-section atlas exists precisely when the complete model space is -finite. Exact functors from finite model semantics, together with their lift spaces, satisfy finite-stage descent, while explicit profinite examples separate objectwise trivialization from uniform parameter elimination.
math.LO — Logic
We show that a Michael space exists in the Laver model.
math.NT — Number Theory
We prove that the modular degree of every elliptic curve with is divisible by .
math.CO — Combinatorics
We prove that the cluster monomials of every graph Laurent phenomenon algebra form an atomic positive basis. For every finite normalized Laurent phenomenon algebra over a unique factorization domain, we construct a simple lattice polytope with unimodular normal fan whose polar realizes the cluster complex. We also prove that cluster monomials form a basis of the upper algebra in the finite case. In particular, every linear Laurent phenomenon algebra admits such a polyhedral realization.
math.LO — Logic
We study an explicit 51-state, 2-symbol beeping Turing machine and its associated dispatch-level process . An ordinal descent below proves that every battle of terminates and hence that beeps infinitely often. A growth lemma shows that the battle counters and beep times eventually dominate every function provably computable in Peano Arithmetic. Together with a Rocq-checked simulation theorem, these results establish that quasihalting of is independent of Peano Arithmetic. The document includes the process definition, machine table, ordinal analysis, growth proof and an account of its external dependencies.
math.FA — Functional Analysis
We construct a symmetric locally finite set Λ ⊂ R for which the space MΛ of crystalline measures with support and spectrum in Λ has countably infinite Hamel dimension; Meyer asked what the dimension of such spaces can be. More precisely, there is a sequence of crystalline measures such that every element of MΛ is a unique finite linear combination of them. Every layer of MΛ of fixed Hermite–Sobolev order is finite-dimensional, but the layer dimensions are unbounded; in one construction they grow linearly, and in general their growth can be prescribed up to a linear change of the order. (For every locally finite set the dimension is finite, countably infinite or equal to the continuum, and the countable case is exactly this behaviour of the layers.) The construction uses clusters of atoms that imitate derivatives of Dirac masses, which produce one-dimensional blocks of prescribed Hermite order. The main difficulty is to show that the union of infinitely many blocks supports no other crystalline measures. We prove this by placing the blocks at random scales, separated by rapidly growing gaps, and showing that every element of fixed Hermite order splits into finitely many block components. We also obtain disjoint sets A, B such that the crystalline measures with support in B and spectrum in A form a countably infinite-dimensional space, while MA = MB = {0}. The existence statement, without the symmetry of Λ, has been formalised and machine-checked in Lean 4; the formal development is recorded in Palomar
math.QA — Quantum Algebra
We give a further combinatorial proof of Laurent positivity for skew-symmetric quantum cluster algebras.
math.GR — Group Theory
A group has property if every automorphism has infinitely many twisted conjugacy classes. Let be an irreducible spherical Artin group of Coxeter rank at least two, with pure subgroup . We classify property for the crystallographic quotient . If has Coxeter rank at least three, then has . In type , this holds exactly when , or when is odd and . Together with the known dihedral Artin-group cases, the positive quotient results imply that every irreducible spherical Artin group of rank at least two has . The positive proofs combine the rational representation theory of the holonomy action with integral lattice constraints and explicit fixed-vector witnesses. In the negative dihedral cases, we construct automorphisms compatible with the crystallographic extension. For type , we exhibit an automorphism with Reidemeister number , showing concretely that the quotient construction does not preserve . This example, including the identification with the presented Artin quotient and the exact Reidemeister count, is formalised in Lean. An alternative proof for seven Artin groups is formalised under explicitly stated external hypotheses.
math.GN — General Topology
In the 2020s, Marzantowicz and Michalak have introduced the notion of system of smooth hypersurfaces and discussed Morse functions compatible with them: the hypersurfaces are regular level sets of the functions. This is a kind of fundamental theory on Morse functions and related differential topology and geometric topology. We extend this to a certain case: we discuss systems of regular level sets and critical level sets (regular fibers and singular fibers) of smooth functions with isolated critical values.
math.NT — Number Theory
Let be the sum of the divisors of , and assume the Riemann hypothesis. Ramanujan showed that , where is a sum over the zeros of . Nicolas made this effective, placed the coefficient of in at colossally abundant numbers, and asked for its exact asymptotic value. We show that this coefficient is with , where oscillates in , . The next coefficient is the constant , up to an oscillation of amplitude at most . In particular, the rescaled residual studied numerically by Leshchiner, Musin and Thompson does not converge, and it exceeds at every colossally abundant number with parameter ; this answers their Questions 3 and 4, the second in the negative. We also prove explicit two-sided bounds for the remainder at every colossally abundant number above Nicolas's threshold , where . This part is computer-assisted: rigorous interval arithmetic covers every real parameter , and an analytic argument covers . As corollaries, for all , improving Nicolas's constant 0.582; Ramanujan's limsup is approached from above at every colossally abundant number with ; and an odd-integer analogue holds with an essentially sharp constant. The proof reads colossally abundant numbers as ground states of a free gas of prime modes, in which splits into a prime-square term, a saturation-deficit term and a shell-energy term. All enclosures are checked against exact values computed from a sieve to .
math.LO — Logic
Restriction to a wide subgroupoid of a model groupoid can introduce invariant relations, maps, and objects. For open topological groupoids, a map between restricted objects lifts exactly when its graph is invariant under the full groupoid, and a specified quotient presentation lifts exactly when its kernel relation is invariant. Under coherent restriction, these criteria recover definable maps and coherent quotient presentations. For finite continuous actions of a group with an open subgroup of finite index, coinduction identifies extensions with invariant subsets on which evaluation is bijective. For a finite group and a normal subgroup , we classify extensions of finite -sets by normalizer extensions of their orbit types. Multiplicities must be constant on conjugacy orbits of types; within each such block, the liftable multiplicities form the additive monoid generated by indices of subgroups over which the normalizer extension splits. The same subgroup data determine the groupoid of lifts and its isomorphism-class and mass generating functions. This classification also applies to finite continuous actions of profinite groups with an open normal subgroup. For a fixed action, a centralizer extension gives the splitting obstruction, with an obstruction and an classification when its kernel is abelian. The center of gives the monoid ; Heisenberg groups over fields of odd prime order give unbounded lifting indices at fixed cohomological period. Finite torsor theories and infinite pure sets realize the definability criteria. For profinite extensions with finite kernel, one finite quotient detects all liftable multiplicities, though it need not carry every lift.
math.NT — Number Theory
We prove lambda_SSS >= 9011123/5000000 = 1.8022246 for the smallest limiting trace-to-degree ratio of totally positive algebraic integers. The previous best bound is 1.8022225 (Tomar). The certificate consists of fourteen integer polynomials with rational multipliers and a continuous piecewise-linear density with 8678 rational hat weights; its dual inequality is verified on (0,infinity) in ball arithmetic. For the same family of polynomials the value of the underlying optimisation problem is 1.8022247 to within 10^-9 (numerical), so the bound is within 1.1*10^-7 of what this family can give.
math.AG — Algebraic Geometry
We prove the Ideal Generation Conjecture for general reduced points in projective space over an algebraically closed field of characteristic zero. Equivalently, if the ideal of the points has initial degree D, multiplication by linear forms from degree D to degree D + 1 has maximal rank. The proof combines hyperplane specialization with a construction using two transverse hyperplanes. In the latter construction, motion of their common trace gives the required rank. A uniform induction for all degrees D ≥ 4 reduces the theorem to explicit quadratic and cubic constructions.
math.PR — Probability
We prove an exact moment criterion for survival in a binary growth process subject to independently renewed spatial catastrophes. Occupied sites spread through a finite set of allowed increments on the integer lattice, and Poisson catastrophes with unbounded cube sizes empty all occupied sites in each affected cube. Under finite one-site hitting intensity, a positive finite-seed survival phase exists if and only if the cube-size measure has a finite moment of order one more than the spatial dimension. The divergent endpoint yields extinction at every positive intensity. The proof centres on a one-dimensional positive-phase construction for connected interval errors under a finite second moment. Extremal-path resampling gives a conditional correlation inequality, which controls exact interval insertion counts and a finite-history potential. Coordinate restriction and envelope covering then establish the criterion in every dimension. We also prove the bulk-to-seed reversal identity and extend the criterion to uniformly fat shapes. For the one-sided one-dimensional rule, the stationary hole density has a quadratic dilute expansion under the same second-moment condition. Its excess over independently eroded Poisson events is strictly positive at every positive intensity, with an explicit quadratic coefficient in the dilute limit.
math.PR — Probability
Take two independent samples of uniform points in the unit square. We prove that, with probability at least one half, the points can be paired so that the average of the exponential of the fourth power of the normalized pair distance is at most two. The distance normalization is , with an absolute constant. One pairing therefore satisfies both coordinate inequalities in the symmetric case of Talagrand's matching problem. The argument first controls a vector field transporting smoothed point densities, and then converts that control to a pairing. The main estimate comes from a Gaussian construction whose successive partitions depend only on noise already revealed; a correction using the classical bottleneck matching preserves the required density balance.
cs.FL — Formal Languages and Automata Theory
For every fixed Boolean grammar in the standard language family considered by Okhotin, we give a deterministic membership algorithm using working space on an input of length . In particular, the whole family is contained in , answering Okhotin's Problem 4 affirmatively. The construction partitions substring endpoints into blocks of size . It represents the parsing recurrence by a constant-fan-in circuit with -bit values and dependency height . Splits in the endpoint blocks are resolved locally, and an ordering of the remaining splits controls the height even when a split is highly unbalanced. Applying the Cook–Mertz tree-evaluation algorithm to the implicitly unfolded circuit gives the space bound.
math.DS — Dynamical Systems
We prove the two specified-phase orbit statements in Shen's Conjectures 3.4 and 3.7. The all-ones phase for the alphabet , with positive and even, and both alternating phases for , with any even integer, have orbit length at depth . The proof uses unreduced signed words and a palindrome class preserved by taking the section of the square at two phase coordinates. It also gives the all-ones statement for nonpositive even under Shen's displayed signed recursion. These are finite phase-orbit results; they do not imply the density conjecture for the classical Kolakoski sequence.
math.DS — Dynamical Systems
Using the frequency coding of Jamet, Marcovici, Poirier, and de la Rue, we determine the largest frequency of in a smooth sequence over and the largest such frequency below . They are and , respectively. The unique attaining type directives are and . The proof is an order comparison for two planar homographies. The first value corrects the numerical expression in the stated extremum question.
math.RA — Rings and Algebras
We disprove the subtraction-free conjecture of Boocher and Froehle for bounded ratios of minors of totally positive matrices. We construct products (P,M,Q) of ordinary minors of a matrix of order (3207) such that [ 0<\frac{P}{MQ}\leq\frac1{64} ] throughout the totally positive locus, whereas (MQ-P) has a coefficient equal to (-1) in the standard bidiagonal parameters.
math.NT — Number Theory
We prove that , a value of Legendre's chi function at a unit of norm , is irrational for every integer except . For this shows that is irrational, where is the golden ratio. The method extends to all units of norm in real quadratic fields, giving outside an explicit set of units, and, through a twist identity, to the inverse tangent integral at units of norm , outside units. The proof combines Hankel determinants of a reflection-symmetric Padé machine, arithmetic bounds on their denominators, a discrete log-gas bound certified in interval arithmetic, and a nonvanishing argument modulo a prime. Below the range covered by the criteria of David, Hirata-Kohno and Kawashima, these results appear to be new. All three theorems are formalised in Lean 4 with Mathlib, using only the standard axioms.
math.AT — Algebraic Topology
We study the projection geometry induced by a singular point on the boundary of a three-dimensional fan-shaped cylinder. The cylinder is defined in the first quadrant of the real coordinate plane, with x and y both between zero and one, and a singular point located at x=1 on the boundary. Through an iterative projection process involving the angle bisector of the quadrant, we prove that the continuous interval [0,1] is precisely discretized to an infinite set of boundary points.
math.CO — Combinatorics
For a graph and a positive integer , the -recolouring graph has the proper -colourings of as vertices, two of them adjacent when they differ at exactly one vertex. A graph is -degenerate if each of its non-empty subgraphs has a vertex of degree at most . The Cereceda (2007) conjecture states that, for fixed and , the diameter of is for every -degenerate graph on vertices. For , the conjecture was known to hold for by a result of Cereceda; for , the best bound known for all -degenerate graphs is , due to Bousquet and Heinrich. We prove that the diameter of is at most for every bipartite -degenerate graph on vertices; hence the conjecture holds for these graphs. We also show that the order cannot be improved for this class, even for graphs of girth . For the upper bound, we fix an ordering of the vertices in which each vertex has at most two later neighbours. In stages, we transform any proper -colouring into the colouring in which every vertex has the least colour not used by its later neighbours, recolouring each vertex at most twice per stage. Both of our bounds have also been proved in the Lean proof assistant.
math.HO — History and Overview
The Grothendieckian Programme is a set of seven works on machine reasoning about mathematics (34 volumes, 26,021 pages), released openly in October 2026 as first drafts that have not been peer-reviewed. This paper draws from them twelve theses on how mathematics and machine reasoning should be combined. Each thesis is stated in one sentence, argued, tied to the passages of the seven works on which it rests, related to existing work, and paired with an observation that would refute it. The theses concern the unit of mathematical knowledge for a reasoning system, the place of epistemic status, the latent space of mathematical reasoning, attention, failure, authority, and the method of discovery. The paper adds no mathematical result to the seven works; none of them claims a proof of the Riemann Hypothesis, the Birch and Swinnerton-Dyer conjecture, the Hodge conjecture, the abc conjecture or the Langlands correspondence.
math.CV — Complex Variables
We prove that Smale's mean value constant exceeds , improving the bound of Jatar and Ng. Smale proved in 1981 that for every complex polynomial of degree at least and every point with , some critical point of satisfies with , and he conjectured that suffices. The best constant is called Smale's mean value constant. Smale's conjecture is false: a proof that appeared in 2026. Jatar and Ng showed that every pair of a normalized univalent map of the unit disk and a finite Blaschke product with , such that has no zero in the disk, gives . We construct such a pair explicitly, with of degree two, from a polynomial of degree with Gaussian rational coefficients, and we verify the required properties by a computation in exact arithmetic. Since the best constant for degree is nondecreasing in (Crane), also exceeds for all large . All theorems, lemmas and corollaries of this paper, and the results of Jatar and Ng that it uses, are also proved in the Lean 4 proof assistant and checked with the Lean comparator.
math.NT — Number Theory
We prove that the closed Báez–Duarte subspace of the Mellin step space is exactly the common kernel of the derivative evaluations prescribed by the zeros of the Riemann zeta function in , with their multiplicities. No finiteness or sparsity assumption on this zero set is required. The proof combines a uniform fractional regularity estimate for Mellin transforms of step functions with a comparison obtained by reflecting zeros in a shifted vertical line. These estimates produce a holomorphic Hardy-space family whose membership in the Báez–Duarte subspace continues from large positive shifts to every sufficiently small positive shift. A uniformly bounded weak limit then gives the synthesis theorem.
math.NT — Number Theory
We prove linear independence results for periods of CM abelian varieties and Beta values by using a known relation between periods as the period term in the algebraic subgroup theorem. For example, , , and are linearly independent over , and families of such relations attached to the CM types of give the linear independence of , a common ratio and periods. The same method proves Waldschmidt's Conjecture 42 at six classes of CM ratios (for instance, and are not both algebraic on the lemniscatic curve) and several four exponentials statements at CM periods. All results are explicit consequences of theorems of Waldschmidt and of Roy and Waldschmidt.
math.QA — Quantum Algebra
Let be the Frobenius–Perron dimension of a weakly integral fusion category and its square-full part. For a divisor-minimal braided obstruction to weak group-theoreticity, we bound the conductors of character fields by formal codegrees and relate residue blocks to primary twists through local Weyl operators. For primes these restrictions give weak group-theoreticity whenever divides or , with arbitrary. Fixing the cofactor of a prime power gives a finite test even when its exponent is unrestricted. A braided category is weakly group-theoretical for and solvable for odd . The corresponding bounds without a braiding are and odd . Combining the dimension families with an exhaustive degree search gives rank at least for a non-pointed simple integral modular category, at least in nonsquare dimension, in odd dimension and in odd nonsquare dimension. We also obtain rank bounds without simplicity and prime-divisor criteria for a braided product decomposition.
math.NT — Number Theory
We prove that the modular curve of the normalizer of a non-split Cartan subgroup of level 121, of genus 511, has exactly seven rational points, all of them CM. This is the level 121 case of a conjecture of Rouse, Sutherland and Zureick-Brown, and with their theorem it determines the 11-adic images of Galois of non-CM elliptic curves over . The Jacobian has rank equal to the genus. We work at the primes 17 and 43, at each of which one Hecke block of the Mordell–Weil group has logarithm divisible by while its -adic height has valuation exactly one. A comparison of valuations in the quadratic Chabauty identity then shows that the residue disc of each CM point holds no other rational point, with no computation of the local height at , no basis of the Mordell–Weil group and no model of the curve. The inputs come from expansions of differentials at CM points and from a formula we prove for intersection numbers of CM divisors under Hecke correspondences on , at every prime , and at when , and .
math.CO — Combinatorics
In a complete graph whose edges have labels 1,...,ell, an edge is critical for a matching if its label is smaller than that of a matching edge sharing an endpoint. The parameter gamma(ell) is the asymptotic worst-case, over labelings, of the smallest critical-edge count obtainable by choosing the matching, normalized by the number of pairs of covered vertices; gamma'(ell) is its perfect-matching version. These parameters arise in bounds for finding cliques in random graphs by edge queries in a prescribed number of rounds. We prove 0.419772820486... <= gamma'(4) <= gamma(4) <= 0.4334035. For every ell >= 2 we give an explicit construction proving gamma'(ell) >= 1/2-1/[4(ell-1)]. We also improve the state-of-the-art upper bounds for general ell, obtaining gamma(ell) <= 1/2-4/(5*2^ell) +O(2^(-3*ell/2)) as ell tends to infinity, together with explicit finite bounds. The proofs of upper bounds compare matchings obtained by local exchanges. The proofs of lower bounds use labelings that are constant on pairs of vertex parts; the stronger four-label construction has six parts and an algebraic value of degree six. Exact-arithmetic verification programs and their data are included for the computer-assisted parts of the proofs.
math.CO — Combinatorics
Let be the number of edges incident to two bounded triangular faces in a finite affine arrangement of pseudolines, with each pair crossing exactly once and with only double and triple crossings. If is the number of triple crossings, we prove , answering the compensation question of Parpalak and Utkin. The proof distributes charge along the normalized boundary of the union of triangular faces. A double crossing incident to four triangles is treated by restricting the arrangement to the four pseudolines supporting their opposite sides. This restriction provides two boundary edges whose unused endpoint charges pay for the crossing.
math-ph — Mathematical Physics
We show that a valid holographic entropy inequality can lack a proof by contraction until all its coefficients are multiplied by a common integer, even when it is balanced, that is, every party has the same total coefficient on both sides, or superbalanced, that is, this remains true after any party is exchanged with the purifier. Every facet of the holographic entropy cone other than subadditivity is superbalanced. A proof by contraction is a map between Boolean cubes that respects the terms in which each party occurs and does not increase weighted Hamming distances. A balanced seven-party inequality and a superbalanced six-party inequality fail at a single point of the cube. Six of their right-hand side terms correspond to the edges of two disjoint triangles of parties, and at that point an unscaled map would have to choose three of these edges that together touch all six vertices, which is impossible. The seven-party one has a proof after doubling. Superbalanced eight- and seven-party inequalities have no unscaled proof even on their realizable domains, the subsets of the cube on which a map already proves validity, and they keep this property when instances of the monogamy of mutual information are subtracted in five successive steps. Splitting further into valid inequalities leads to a superbalanced facet of the seven-party cone that we did not find in the literature. It has unscaled maps on its realizable domain, all of which break its symmetry. For star inequalities, whose left-hand side contains only single-party entropies, the least scaling is the least for which an explicit polytope contains a point with coordinates in . From the weighing matrices of Alon and V\~u we obtain valid -party star inequalities whose least scaling exceeds for every and large , while every finite least scaling of an -party star inequality is below . Non-existence rests on DRAT proofs, validity on exact rational maps or identities, and the facet property on explicit graph models, except for the star inequalities of arbitrary size, whose results rest on written proofs.
math.PR — Probability
Benjamini and Schramm conjectured in 1996 that site percolation on every nonamenable quasi-transitive graph has . We prove this in Hutchcroft's operator form: for Bernoulli site percolation on every connected, locally finite, nonamenable quasi-transitive graph, the critical two-point operator is bounded on , , and in the standard monotone coupling infinitely many infinite clusters coexist at every parameter of a nondegenerate interval. Since bond percolation is site percolation on the line graph, this contains OpenAI's recent bond theorem. For unimodular automorphism groups we adapt OpenAI's proof: cut vertices replace bridges, a pruning lemma handles retained contact vertices, a rejection step controls the junctions created by opening a vertex, and pivotal corridors become induced paths. For nonunimodular groups we extend Tran and Nguyen's site argument from vertex-transitive to quasi-transitive graphs.
math.DS — Dynamical Systems
Boshernitzan conjectured in 1992 that in a rational right or isosceles triangle every direction of an explicit finite set is purely periodic: every billiard orbit in it that avoids the vertices is periodic. For a right triangle with acute angle in lowest terms the set is two classes of directions, each containing the perpendicular to a side, and the conjecture says that both are completely periodic. It fails at one known right triangle, . We prove it, when , whenever and are both prime to , a set of density about , and whenever has at most one odd prime factor, with a computer-assisted check of finitely many cases; and, through the fold, for the corresponding isosceles triangles. Everything turns on one object, an orbit from a vertex to a vertex that never strikes a side at a right angle: if there is none, both classes are completely periodic, and at odd the retracing band of a companion paper exists. Such an orbit satisfies an identity among roots of unity whose coefficients have signs fixed by the geometry, and Galois conjugation — or, at a single odd prime, a slowest-rotation argument — makes it impossible. When neither nor is a unit modulo , no argument from these inputs can succeed, and the counterexample lies there. A second method reads the poles of a minimal component as a word of a binary code attached to ; it reaches further composite and, with interval arithmetic, the triangles at every . Beyond right triangles we prove, with computer assistance, that Boshernitzan's Conjecture 2.2(a) fails for infinitely many triangles.
cs.CG — Computational Geometry
We study how low-frequency structure, measured with respect to the Laplacian of a fixed graph built on the training samples, propagates through a ReLU network to the least-squares readout , and how individual neurons can be optimized to capture low-frequency eigenvectors. The readout error splits into a low-frequency part controlled by coverage (how well the feature span contains the low eigenvectors) and a high-frequency part controlled by the smoothness (Dirichlet energy) of the target. We prove an error bound that makes this precise, with coverage measured by the coverage gap, the largest principal-angle sine between the low eigenvectors and the feature span, and show that coverage can always be preserved across ReLU layers. Finally, we define a count of captured low-frequency features, relate it to our error bounds, and prove that error can be reduced to a smoothness component when captures every low frequency feature. Lastly, we show that a ReLU layer can strictly increase whenever the previous layer keeps the samples distinct and coverage is not yet full, that neurons suffice to preserve coverage, that greedily chosen neurons drive the coverage gap down at rate governed by a ReLU variation norm of the low eigenvectors, and that neurons can reduce the coverage gap no further than the -st principal-angle sine.
math.CO — Combinatorics
We prove that by constructing a red/blue coloring of the edges of with no monochromatic . We start from a classical -coloring on 1596 vertices, delete 21 of its vertices and add 85 new ones. Toward the remaining 1575 vertices, each new vertex has either the same colors as one of the deleted vertices or a single color. The deleted vertices include two pairs whose members have complementary colors toward every other vertex, and these pairs weaken the conditions that the new vertices must satisfy. The proof reduces to a finite check on the new vertices and is formally verified in Lean 4.
math.NT — Number Theory
Let be a primitive element of and, for , let be the corresponding codeword of the dual Melas code. Let be the number of occurrences of a uniformly random pattern in . Sac Himelfarb and Schwartz conjectured that converges to the Poisson law for every choice of ; they proved that the second factorial moment satisfies , and for almost every sequence . We prove that uniformly in and in the primitive element . Thus the second factorial moment converges to for every sequence . The proof expresses through a norm form of the quadratic extension , which turns it into a bilinear Kloosterman sum of full length, and bounds the latter in characteristic two. We also show that the counts , as functions of , are the fibre sizes of a family of quadratic maps, and deduce that for almost every sequence of patterns the first two factorial moments of converge to , as they would under the Poisson law conjectured by Sac Himelfarb and Schwartz for fixed patterns.
math.DG — Differential Geometry
This article presents seven AI-generated research directions in differential geometry and differential topology, each centered on a proposed research object or formulation. For each direction, we provide basic definitions and theorems, discuss connections with existing mathematical problems and theories, and outline questions for further study. The definitions introduced in these seven directions are relatively new. The article is intended as a collection of AI-generated research guides rather than a conventional research paper.
math.AG — Algebraic Geometry
This is a report on two preprints of OpenAI, " Milne's rationality conjecture for abelian varieties" and "The rational Hodge conjecture for CM abelian varieties", placing them in context.
cs.CC — Computational Complexity
We describe a natural new example of non-computability in analysis. Namely, the multiplicative autocorrelation of the fractional-part function, \[ A(x)=\int_0^\infty \{t\}\{xt\}\,\frac{dt}{t^2}, \] is computable and differentiable almost everywhere, but its finite derivative can be non-computable even for a computable . We derive these results from an adaptation of our earlier work on computability of Brjuno-like functions.
math.CO — Combinatorics
For a finite cyclic group and an exact Fourier support , we study the lattice generated by zero-sum relations of length at most inside the full relation lattice . Its closing degree is both a local separating degree and the order at which every invisible phase ratio becomes a translation character. For the exact unit support of every even cyclic group, we compute $$ \mathbb{Z}^{U_{n}}/\mathcal{L}_{\le4}(U_{n})\cong \begin{cases} \mathbb{Z}/n\mathbb{Z},&3\nmid n,\\ \mathbb{Z}\oplus\mathbb{Z}/(n/6)\mathbb{Z},&6\mid n. \end{cases} $$ Moreover . Thus fourth- and fifth-order rational ambiguity coincide: fourth order reconstructs up to translation when , while for the complete surviving family has normalized ratios parameterized by norm-one elements of , modulo the residual translation subgroup ; order six closes it. For with distinct odd primes, we classify every nonempty -stable support. The distributions of these 255 supports by closing degree are $$ (1,2,154,80,4,14)\quad(p=3),\qquad (1,2,154,92,4,2)\quad(p,q\ge5). $$ The proof uses seven local CRT presentations, cubic and quartic Skeletons, and explicit depth-five and depth-six obstructions. Galois-equivariant lower-bound ratios show that the lattice, rational-signal and integer-signal closing degrees coincide on every stable support. Finally, for binary indicators, a supportwise crosswalk and a uniform exclusion of all deep supports give under translations. The value is also four for the same oriented analytical deck when outputs are identified up to translation or reflected translation.
math.CO — Combinatorics
Let be a finite alphabet with a fixed-point-free involutive complement, and let reverse complement act on the directed de Bruijn graph by reversing a word and complementing every letter. We prove that a reverse-complement-invariant ordinary minimum decycling set exists if and only if is even, for every integer . This simultaneously covers the binary alphabet and the DNA alphabet with and . For odd , a primitive rotation cycle joins two reverse-complement-fixed lower states in opposite directions. Disjoint complementary alphabet pairs force at least deletions beyond the ordinary minimum. For even , choose injective antisymmetric real weights on the alphabet and form a weighted cyclotomic selector. We prove that arbitrary representatives on its zero-embedding rotation cycles preserve acyclicity, resolve those ties equivariantly, and establish the exact defect identity A finite sequence of legal full-alphabet source firings removes this nonnegative symmetric defect two coordinates at a time. We also identify a self-dual spectral ideal that gives a deterministic direct selector, rather than a simulated firing path. For integer weights, membership is decidable with polynomial bit complexity. The executable finite panel in the repository checks the binary specialization only; it is convention-level assurance and is not used in the all-order proof.
math.QA — Quantum Algebra
We give another proof of Laurent positivity for skew-symmetric quantum cluster algebras.
math.AC — Commutative Algebra
We prove Laurent positivity in the original exchange coefficients for skew-symmetrizable Chekhov–Shapiro generalized cluster algebras of arbitrary finite rank.
math.HO — History and Overview
The Grothendieckian Programme is a series of seven works on machine reasoning about mathematics (34 volumes, 26,021 pages), released openly on Zenodo in October 2026 with an overview volume. Two of the works specify a reasoning architecture — a core that holds mathematical worlds as typed, versioned and verifiable objects, and a kernel that reasons inside such worlds by sites, sheaves and descent. Five apply one method, recursive mathematical discovery, to the Riemann Hypothesis, the conjecture of Birch and Swinnerton-Dyer, the Hodge conjecture, the abc conjecture and the Langlands programme, and record what it proves, what it reduces to stated hypotheses and what it leaves open. This paper describes the architecture, the five research subjects, the method, the published corpus and the open process by which the works are to be reviewed and continued. None of the works claims a proof of any of these conjectures; they are first drafts and have not been peer-reviewed.
math.CO — Combinatorics
We determine the least pool sizes guaranteeing a frame or a complete box from unit cubes whose faces use all five colours of a fixed palette, for every . Proper rotations are allowed; the target has monochromatic exterior faces using all five colours. The frame threshold is , and the complete-box threshold is . The proof combines a certified fifteen-cube corner theorem with elementary Hall arguments and a uniform three-variety obstruction. It completes the five-colour frame column left open by Berkove, Eickemeyer and Schulman (2018). All threshold statements are formalised in Lean. A fixed-size flow formulation decides feasibility and constructs assemblies in time linear in the explicit source list. We also determine the corner thresholds when the number of source varieties is bounded, and classify all fourteen-cube obstructions into 56 classes under common colour relabelling and simultaneous reflection, described by three structural forms. The minimum size of a collection capable of reproducing every target is also fifteen.
cs.FL — Formal Languages and Automata Theory
For every fixed finite alphabet with at least two letters, we construct smooth decomposable circuits representing the primitive words of each length . The gates are literals, Cartesian products on disjoint position sets, and disjoint unions on equal position sets. The total number of gates and wires is , and the construction takes bit operations. Prime periods are expressed as coordinate directions of independent rectangular grids. A recursion at the first nonconstant vertex computes the exact set of failed directions; sharing its continuations gives the size bound. This resolves the primitive-word instance of the circuit question posed by Amarilli, Monet, Raphael, and Salvati. It does not determine whether primitive words form a context-free or a poly-slicewise-unambiguous language. The same circuit bound holds for any union of exact power orders at a fixed length.
math.NT — Number Theory
We determine all the groups that appear as the torsion group of an elliptic curve over a number field of degree . There are such groups. For of them, namely , , and , we also determine all the elliptic curves over number fields of degree that realize them, and there are only finitely many. These four groups were not known to occur over number fields of degree . The proofs combine the global method of Derickx, Hecke sieves, formal immersion criteria at the prime , and Riemann-Roch computations on modular curves and their quotients over finite fields.
math.CO — Combinatorics
We prove Cazenave's conjecture for connected graphs, where and are the second largest adjacency and Laplacian eigenvalues. We also characterize equality in this inequality and in the linear bound .
math.NT — Number Theory
We show that the modular curve , attached to the normalizer of a non-split Cartan subgroup of , has exactly rational points, all of them CM points. Together with the work of Rouse, Sutherland and Zureick-Brown, this completes the classification of the possible -adic images of Galois of elliptic curves over without complex multiplication. We do not use a model of , which has genus . Instead we use the map , whose target is an elliptic curve of rank , a descent with the cusps of and a Mordell-Weil sieve to reduce the problem to hyperelliptic curves of genus over a number field of degree . We deal with these curves by a version of Stoll's Selmer group Chabauty method for the isogeny of their Jacobians.