Mathematics

AI contributions across the mathematics fields in the catalog, including algebra, analysis, probability, combinatorics, geometry, topology, number theory, algorithms, optimization, and formal proof.

Curated contributions 984

Understand the topics ↓

Contribution counts by ratings

AI involvement:
AI-centered250AI-led646AI-autonomous88
Importance:
Modest76Notable509Major399
Technical verification:
Reported346Partial106Formal532
Human understanding:
Few629Partial17Full338
How we assess contributions →

Every two-coloring of the plane misses at most one triangle, which is equilateral

  • Mathematics

In every arbitrary two-coloring of the Euclidean plane, each non-equilateral three-point set, including three distinct collinear points, has a monochromatic congruent copy. At most one equilateral side length can be missed, and an alternating-strip coloring shows that one exception can occur.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

Erdős Problem 111: bipartizing finite subgraphs of uncountably chromatic graphs

  • Mathematics

If a graph is not countably colorable, h_G(n)/n tends to infinity, where h_G(n) is the largest minimum number of edge deletions needed to make an n-vertex subgraph bipartite. Equivalently, a uniform linear deletion bound on all finite subgraphs forces countable colorability. The limit holds for the whole sequence, for graphs of any uncountable chromatic cardinality.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

The Selmer converse for elliptic curves at every prime

OpenAI

  • Mathematics

For every elliptic curve A/Q, every prime p and full p-power Selmer corank r in {0,1}, the manuscript claims analytic rank and Mordell–Weil rank r and finiteness of the whole Tate–Shafarevich group, without reduction, CM, residual-representation, torsion or isogeny restrictions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Persistent failure of Euclidean quartic SOS exactness

  • Mathematics
  • Theoretical Computer Science

An explicit fixed rational cubic in 14 variables gives a quartically regularized polynomial whose exact minimum shift is nonnegative but not a sum of polynomial squares for every positive regularization parameter. A second explicit cubic has a unique global minimizer with positive-definite Hessian there and retains a strict SOS gap for every positive parameter.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

A Limiting Satisfiability Threshold for Every Fixed Clause Size

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every fixed k≥3, random k-SAT with independent uniformly signed proper clauses sampled with replacement has a positive finite limiting density αₖ: satisfiability probability tends to one below αₖ and zero above it. The claim makes no assertion at αₖ.

AI involvementAI-led
ImportanceModest
Technical verificationFormal (Lean 4 with Mathlib)
Human understandingFew

Hyperbolicity Cones Without Semidefinite Lifts

OpenAI

  • Mathematics
  • Theoretical Computer Science

An OpenAI manuscript claims a real homogeneous hyperbolic polynomial whose closed cone has no exact finite affine semidefinite lift, even with arbitrary real coefficients and any finite number of auxiliary variables.

AI involvementAI-led
ImportanceMajor
Technical verificationPartial
Human understandingFew

Exact mass asymptotics for the two-dimensional O(4) lattice model

OpenAI

  • Mathematics
  • Theoretical Physics

For the nearest-neighbor square-lattice O(4) model, the full Osterwalder–Schrader transfer gap, including rotation-invariant local-observable sectors and measured per original lattice time step, is asymptotic to 32 exp(π/4−1/2) √β exp(−πβ) as β tends to infinity.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

The Bollobás–Nikiforov inequality for nonnegative edge weights

Gabriel Coutinho, Yinchen Liu, Thomás Jung Spier, Quanyu Tang, Shengtong Zhang

  • Mathematics
  • Theoretical Computer Science

The authors prove λ₁(G)²+λ₂(G)²≤2(1−1/ω(G))|E(G)| for every finite simple noncomplete graph G with at least two vertices. A stronger weighted theorem bounds the two largest positive eigenvalues of a symmetric nonnegative zero-diagonal matrix supported on G by (1−1/ω(G)) times its squared Frobenius norm.

AI involvementAI-centered
ImportanceMajor
Human understandingFew

A Smooth Metric with No Local Isometric Immersion into Three-Space

OpenAI

  • Mathematics
  • Theoretical Computer Science

OpenAI’s Theorem A constructs a C∞ positive-definite metric g on (-1,1)^2 such that no open neighborhood of the origin admits a C∞ isometric immersion into R³. The paper also says g can agree with the Euclidean metric to every order at the origin. The Lean endpoint directly states the smooth positive metric and no-local-immersion claim; the flat-jet refinement is not part of that endpoint.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

AI agents extend a difference-free construction to all powers

  • Mathematics
  • Theoretical Computer Science

For every k≥2, a nonempty pool of ranked blocks on pairwise coprime square-free moduli, with ranks strictly decreasing along nonzero k-th-power-residue difference arcs, gives α(k,P)≤liminf log D_k(N)/log N and eventually D_k(N)≥N^ρ for every ρ<α(k,P). Two separate nine-block instances give α(4,P₄)=0.912145042700…>0.9121 and α(6,P₆)=0.950825559086…>0.9508; each exceeds the specified six-residue Ruzsa-transfer comparator.

AI-led Lean proof settles a seven-vertex unit-distance case

  • Mathematics
  • Theoretical Computer Science

For every simple graph G on seven vertices, either G or its complement has an injective unit-distance representation in Euclidean R³, where every edge has length one and nonedges are unconstrained. This proves the d=3 case of Frankl–Kupavskii–Swanepoel Problem 2; their general 2d+1-vertex question remains open. Combining with the published lower bound gives f_D(7)=3, although the lower bound is outside the formal target.

AI involvementAI-led
ImportanceModest
Human understandingFew

Lean proof characterizes two-fold Langford sequences

  • Mathematics
  • Theoretical Computer Science

For d,l≥1, a two-fold Langford sequence of order l and defect d exists exactly when 3l≥2d−1. Each number p in [d,d+l−1] occupies endpoints of two disjoint pairs at distance p among 4l positions; the proof uses sixteen small cells, concatenation and signed-permutation families.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

Determinant Divisibility of Centered Latin Squares: Binary Criteria, Universal Counterexamples, and p-Adic Bounds

Oleksiy Babanskyy

  • Mathematics

For even-order Latin squares, Babanskyy gives binary-rank and adjugate criteria for an extra factor of two in the centered determinant. An explicit family has an odd normalized determinant quotient for every n≡2 mod 4 with n≥6.

AI involvementAI-centered
ImportanceModest
Human understandingFull

Cycles of length divisible by five in graphs of minimum degree five - The k=5 case of Dean's conjecture

Elias Botsford

  • Mathematics
  • Theoretical Computer Science

Botsford’s revised theorem says every finite simple graph with minimum degree at least five has a simple cycle whose length is divisible by five. The proof reduces a minimal counterexample to 5-weak graphs and treats bipartite, triangular and triangle-free cases.

AI involvementAI-led
ImportanceMajor
Technical verificationPartial
Human understandingFew

Depth-Two Smith Profiles and Carry Geometry for Affine Hjelmslev–Radon Incidence

Oleksiy Babanskyy

  • Mathematics

Babanskyy determines the complete p-primary Smith profile of the integer affine point–line incidence matrix over Z/p²Z for every prime p, including the nonunit factors p, p² and p³ and their multiplicities. The argument also gives a nonsplit transfer extension and a two-chart carry presentation.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingPartial

Depth-Three Smith Profiles and Newton-Minor Geometry for Affine Hjelmslev–Radon Incidence

Oleksiy Babanskyy

  • Mathematics

Babanskyy reports every nonzero Smith factor of the integer affine point–line incidence matrix over Z/p³Z for every prime p. The profile reaches exponent p⁵ and has six multiplicities, with polynomial formulas for odd p and explicit p=2,3 rows. The new step resolves a q-Pascal common image through q-Newton minors and a cross-chart carry calculation.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingPartial

AI-centered Lean proof classifies base-three partition counts

  • Mathematics

For partitions of {1,…,n} into nonempty blocks of at most three elements whose sums are powers of three, the count is exactly one on N_u={1,2,3,4}∪{3ᵗ+d:t≥2, d∈{−4,−2,−1,0,1,2,3,5}}; exactly two on E₂={13}∪{3ᵗ−3:t≥2}∪{3ᵗ−6:t≥3}; and at least three otherwise.

Polynomial Hitting Lists for Noncommutative Rational Formulas

OpenAI

  • Mathematics
  • Theoretical Computer Science

Given unary n and a size bound s, a deterministic machine constructs in polynomial bit time a polynomial-size list of rational matrix tuples of polynomial total output length and common polynomial dimension. Every admissible nonzero noncommutative rational formula over Q of size at most s has a defined, invertible evaluation on a listed tuple.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4 with Mathlib)
Human understandingFew

Integral tail signatures and the Hilbert–Smith conjecture

A. Dabrowski's AI agents

  • Mathematics

The selected formal development claims that a continuous additive p-adic-integer action is excluded on a connected, Hausdorff, second-countable, finite-dimensional manifold without boundary. It also claims that a second-countable locally compact Hausdorff group acting continuously and faithfully on such a manifold has a compatible smooth Lie-group structure.

AI involvementAI-led
ImportanceMajor
Technical verificationPartial
Human understandingFew

Quadratic Strong-Operator Paving over Arbitrary Maximal Abelian Subalgebras

OpenAI

  • Mathematics

For every 0<ε<1, every self-adjoint x in any von Neumann algebra M admits strong-operator paving over any maximal abelian subalgebra A with at most 5×10⁸ε⁻² projections. The count is fixed before the inclusion, operator and strong neighborhood, without separability or conditional-expectation assumptions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

AI-assisted Hadamard constructions cover eight orders

  • Mathematics

Using periodic T-matrix quadruples of orders 103 and 163 and symmetric-circulant Williamson quadruples of orders 7, 11, 13, 17, 19, 23 and 29, a formalized Cooper–Wallis construction gives Hadamard matrices at orders 2884, 4532, 7004, 7172, 7828, 8476, 9476 and 11948. A separate parity result rules out Miyamoto’s displayed C2-matrix ingredient at positive block orders not congruent to 1 mod 4.

AI-found residue sets improve a polynomial-free lower bound

  • Mathematics

For positive-integer sets avoiding {x,x+y,x+y²} for every nonzero integer y, Claude-found residue sets modulo 145 of sizes 10 and 32 satisfy Younis’s two-set conditions. They give lower-bound exponent 1/2+log(10)/(3log(145))+log(32)/(6log(145)), between 0.7702 and 0.77029.

AI-built Lean program triangulates simple polygons in linear time

  • Mathematics
  • Theoretical Computer Science

For every closed simple polygon with n vertices, including collinear vertices and equal heights, an explicit exact-real-RAM program returns a conforming triangulation with n−2 triangles in at most C(n+1) steps and natural words bounded by K(n+1). It is a variant of Chazelle’s algorithm that completes the visibility map directly below n=2³⁰+2.

AI involvementAI-led
ImportanceNotable
Human understandingFew

AI-derived determinant restriction for finite-order matrices

  • Mathematics

For m>0 and N≥1, the formalized refinement says GL(N,ℤ) contains an element of order m with determinant −1 exactly when m is even and ψ(m)+1≤N, where ψ sums φ(p^k) over exact prime-power factors except 2. The same criterion holds over ℚ, and the least integral dimension is ψ(m)+1.

AI involvementAI-led
ImportanceModest
Human understandingFew

AI-assisted proof gives sharp graph-matrix norm orders

  • Mathematics

For each fixed finite simple graph shape in the globally injective dense independent-sign model and fixed p>0, the pth norm moment has matching upper and lower orders n^((h+i₀−s)/2)log(2n)^(b*/2), with h the role count, i₀ isolated interior roles, s the minimum boundary-separator size and b* the largest attached floating-component count over minimum separators.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

AI-assisted Walsh embedding reaches linear row scaling

  • Mathematics
  • Theoretical Computer Science

For n=2^k, a fixed real orthonormal n×d frame, 0<ε<1 and 0<δ<1/2, two independent sign rerandomizations of the normalized Walsh transform followed by uniform row sampling preserve all squared norms within ε with probability at least 1−δ. The stated row count is min(n,⌈8192(d+27⌈log₄(4d/δ)⌉³)/ε²⌉), which is O(d/ε²) for fixed δ.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

AI-assisted inverse theorem for short zero sums in C₃³

  • Mathematics

For a length-24 positional sequence of nonzero elements of C₃³, every pair of nonempty zero-sum subsequences of length at most three has zero intersection sum exactly when its positions can be permuted into three copies of eight distinct elements with no short nonempty zero sum.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

AI proof of Gallardo’s all-ones polynomial equivalence

  • Mathematics

For every n≥2 over GF(2), the polynomial 1+X+⋯+X^n is irreducible exactly when its composition with X²+X+1 is irreducible. The formalization also checks the corresponding OEIS set statement and excludes n=1.

AI involvementAI-led
ImportanceModest
Human understandingFew

AI-derived bounds for Mazur’s greedy drawing construction

  • Mathematics
  • Theoretical Computer Science

For a well-formed Mazur edge, cycle, series and join scheme with n≥2 vertices and j nontrivial joins, the span satisfies T+1≤2^j(n−2j), with equality at (n,j)=(2,0) or 2j+3≤n. Its worst case is T=2^(n/2)−1 for even n and T=3·2^((n−3)/2)−1 for odd n. Companion theorems give interior dart mass ≤aₛbₛT, Dirichlet row norm ≤2(n−1)²T and |det M|≤(2(n−1)²T)^n, bound the candidate products T²|det M| and T³|det M|, and prove z=adj(M)u with Mz=Δu and boundary zₛ=Δhₛ(T−hₛ) when hₛ≤T, where Δ=det M.

AI involvementAI-led
ImportanceModest
Human understandingFew

Claude formalizes an alternative Hadamard border argument

  • Mathematics

Claude agents devised an integer proof route for the s=1, index-two clauses of Theorem D in Jones’s Hadamard border note, replacing its rank and positive-semidefiniteness route. The formal development checks those Theorem D clauses.

AI involvementAI-centered
ImportanceModest
Human understandingFew

Astra finds a five-generator semigroup family with exact Betti count

  • Mathematics

For k≥3 and q≥k−1, let m=2kq+1. The numerical semigroup minimally generated by m,m+1,m+k,m+2k−1,m+2k has toric-ideal generator count 4k+2. The formal development also gives a nonredundant presentation.

AI involvementAI-centered
ImportanceModest
Human understandingFew

AI-contributed lemmas help prove square-grid tree maximality

  • Mathematics
  • Theoretical Computer Science

For every n>0 and every set of n² lattice points, the induced square-grid graph has at most as many spanning trees as the n×n square. Equality holds exactly for translates of that square.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

GPT-generated bounds count intermediate subgroups

  • Mathematics

For H≤G of finite index n>1, Palcoux’s manuscript and formal development give a finite rational upper envelope for the number of intermediate subgroups and a strict bound c(p)n^((logₚn)/4), where p is the least prime divisor of n. Selected results also address prime-power and elementary-abelian sharpness.

AI involvementAI-autonomous
ImportanceModest
Human understandingFew

AI-assisted proof establishes fixed Hodge Lie generation

  • Mathematics
  • Theoretical Physics

For diag(1,1,1,−1,1,−1) and the local Hodge operator on coordinate plane 0123, fifteen orthogonal adjoint actions together with that operator generate sl(15,K) over every field of characteristic other than two. The same formal result gives dimension 224 and scalar rigidity for commuting endomorphisms, with determinant c²²⁴.

AI involvementAI-led
ImportanceModest
Human understandingFew

AI-origin Tate-domain examples answer a sheafiness question

  • Mathematics

Birkbeck and Torzewski give uniform, strongly sheafy, nonnoetherian Tate domains that are not stably uniform, answering Kedlaya’s Question 7. One finite-jet example has a nonreduced rational localization; a weighted-parity example has a nonuniform localization that remains a domain.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Claude-led proof gives a logarithmic symmetric-group word threshold

  • Mathematics

For nonzero integers r,s with at least one odd, the word xʳyˢ is universal in Sₙ once n is sufficiently large and at least C log m(r,s), where m is the product of the distinct primes dividing rs. A separate construction shows the logarithmic order cannot generally be reduced.

AI involvementAI-led
ImportanceNotable
Human understandingFew

AI-developed proof improves the permanent-approximation base

  • Mathematics
  • Theoretical Computer Science

For every nonnegative rational n×n matrix, Anari’s formal development constructs one rational-output algorithm and an absolute constant c<√2 with a two-sided cⁿ approximation guarantee for the permanent. It also proves deterministic polynomial time in the ordinary binary input size.

A classification of biologically unavoidable sequences

  • Mathematics

For every non-eventually-periodic sequence over a finite alphabet, the project constructs an infinite labelled population satisfying Alexander’s finite-root, finite-outdegree, finite real-birthdate-sublevel and incoming-label conditions but with no matching infinite path from any start. With Alexander’s earlier positive theorem, biological unavoidability is equivalent to eventual periodicity.

AI-assisted certificate gives the proper hat-guessing number of K₈−e

  • Mathematics

The proper hat-guessing number of K₈−e is 14. A PGL(2,13)-equivariant fourteen-colour strategy uses 990 normalized twin decisions and a 48,510-edge residual-orbit matching; exhaustive checking covers all 138,378,240 proper colourings, and a general upper bound gives equality.

AI involvementAI-centered
ImportanceModest
Technical verificationFormal (exact computation)
Human understandingFew

On A ⊆ [N] Such That ab + 1 Is Never Squarefree for a, b ∈ A

Mehtaab Sawhney

  • Mathematics

For sufficiently large N, Sawhney bounds the size of A⊂[N] with ab+1 nonsquarefree for every a,b∈A by the size of the residue class 7 mod 25, and identifies the extremal 7 and possibly 18 mod 25 classes. The model-origin part is a squarefree-density sieve with prime exclusions depending on b and the progression difference.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

AI-generated certificate candidate for Dittert’s dimension-five inequality

  • Mathematics
  • Theoretical Computer Science

A released AI-generated sum-of-squares and Lean candidate encodes Φ(A)≤1226/625−||A−U₅||²_F/625 for nonnegative 5×5 matrices whose entries sum to five. If its encoding matches the original problem, it implies unique equality at U₅.

AI involvementAI-autonomous
ImportanceModest
Technical verificationPartial
Human understandingFew

A Four-Terminal Planar Lower Bound and a Sharp Equal-Cost Gadget for Cost-Preserving Unsplittable-Flow Rounding

Matthew Protti

  • Mathematics
  • Theoretical Computer Science

Protti’s planar acyclic four-terminal instance needs upper overload at least 335 when the largest demand is 294 if cost cannot increase. A rational family gives a lower bound (299−41√41)/32≈1.139747 for the unrestricted planar constant.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFull

AI-found graph refutes Written on the Wall II Conjecture 109

  • Mathematics
  • Theoretical Computer Science

A connected 13-vertex graph formed by joining an independent seven-vertex set to two disjoint triangles has independence number 7, largest induced bipartite subgraph size 9 and Havel–Hakimi residue 2. These values violate Conjecture 109’s proposed bound of 6.

AI involvementAI-led
ImportanceModest
Human understandingFew

A radial continuum phase transition with algebraic decay

OpenAI

  • Mathematics
  • Theoretical Physics

One bounded continuous stable radial pair potential in three dimensions has |φ(r)|≤C r^(−3−1/32) for r≥1 and finite canonical free energy. At one common βc in [7/8,9/8], the free energy has a strict inverse-temperature derivative jump for every density in an open interval around 5p/3.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean)
Human understandingFew

An exact plateau for three binary summands modulo a Mersenne number, with finite-layer carry balancing

Anonymous

  • Mathematics
  • Theoretical Computer Science

For k≥4, the ordered three-summand count Fₖ(t) modulo 2ᵏ−1, with total binary weight below k, equals (k+23)3^(k−4) when the k-bit target has exactly two nonadjacent zero digits. The paper proves balancing only through carry mass five and does not classify global maximizers.

AI involvementAI-autonomous
ImportanceModest
Technical verificationPartial
Human understandingFew

AI-led proof of Rényi and Tsallis entropy monotonicity for Bernoulli sums

  • Mathematics
  • Theoretical Computer Science

For every finite family of independent Bernoulli variables, increasing each success probability from pᵢ to rᵢ with 0 ≤ pᵢ ≤ rᵢ ≤ 1/2 cannot decrease either the Rényi or Tsallis entropy of their sum, for any fixed entropy order 0 ≤ q ≤ 2. The statement includes zero and fair probabilities, support entropy at q=0 and Shannon entropy at q=1.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

Two Möbius groups suffice: Euclidean Ramsey sets on spheres of every dimension, after the tensor classification

Will Strinz

  • Mathematics
  • Theoretical Computer Science

Strinz reports that every finite set of distinct points on a circle is Euclidean Ramsey when all but at most three points have coordinates algebraic over Q. The circle’s center and radius may be arbitrary real numbers.

Counterexamples to the restricted-sumset inverse theorem in arXiv:2609.10148v2

alexander

  • Mathematics

In ℤ/13ℤ, A={0,1,3,4,5,6,10} and B={0,1,4,6} have |A|+|B|=11≤12. Their restricted sumset is {1,3,4,5,6,7,9,10,11}, of size 9=|A|+|B|−2, but neither A nor B is an arithmetic progression. This contradicts the forward conclusion of Li–Liang Theorem A/3.1 in arXiv:2609.10148v2.

AI involvementAI-led
ImportanceNotable
Human understandingFull

Anthropic reports: Claude reconstructs NIST elliptic-curve seed generation

  • Mathematics
  • Theoretical Computer Science

Anthropic reports recovering a reproducible lrand48-and-timestamp recipe for all five NIST prime-curve seeds, five additional ANSI prime-curve seeds and four NIST binary-curve seeds. The reported timestamps and candidate indices regenerate each 160-bit public seed. The NIST B-163 seed remains unexplained.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFull

Revisiting Separators in String Graphs, Region Intersection Graphs, and Induced-Minor-Free Graphs

James Davies, Paweł Rzążewski, Jack Spalding-Jamieson, Bartosz Walczak

  • Mathematics
  • Theoretical Computer Science

The AI-origin improvement in Appendix B gives non-expansive line embeddings with O(log h) average distortion for the stated vertex-weighted metrics of connected graphs excluding the one-subdivision of K_h as an induced minor. The paper derives balanced separators of size O(h log h √m) for the corresponding m-edge graphs.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFull

AI-assisted sharp support criterion for variance-proxy sub-Gaussian laws

  • Mathematics
  • Theoretical Computer Science

The support-criterion project reports that any symmetric real probability law with 0<a≤|X|≤b almost surely and b/a≤√3+√2 satisfies E exp(tX)≤exp(t² E[X²]/2). The inequality is strict for nonzero t, the variance proxy is optimal, and every larger universal support ratio admits a symmetric four-point counterexample.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

AI-led proof gives a sharp global squared-law transport bound

  • Mathematics
  • Theoretical Computer Science

For any real law with X² in [A,B], A=3−√6 and B=3+√6, let δ=3(E[X²])²−E[X⁴] and let X₀ be the four-node Gaussian quadrature law identified by zero deficit. For every feasible 0≤δ≤30+12√6, W1(L(X²),L(X₀²))≤√(δ/3), using ordinary infimum-over-couplings Wasserstein distance. The bound is attained at every feasible deficit and does not require symmetry of X.

AI-led proof gives an optimal eighth-order MGF gap

  • Mathematics
  • Theoretical Computer Science

For every symmetric law with X² in [A,B], A=3−√6 and B=3+√6, set v=E[X²] and h=E[(X²−A)(B−X²)]. For every real t, exp(vt²/2)−E exp(tX)≥h t⁴/24+t⁸/25200. Both coefficients are optimal for this two-term bound. A positive coefficient on the intrinsic fourth deficit 3v²−E[X⁴] cannot be paired with the optimal eighth-order coefficient.

Discovering state-of-the-art quantum circuits with large language models

Andi Gu, Alek H. Kemeny, Susanne F. Yelin, Mikhail D. Lukin

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

Gu, Kemeny, Yelin and Lukin report lower Toffoli and peak-qubit counts in seven specified circuit benchmarks, plus a new prime-field modular-inversion architecture. For fixed-point Gaussian evaluation, their stated benchmark falls from about 24,000 Toffolis and 362 qubits to 1,583 Toffolis and 139 qubits; the reported worst-case error is 9.99×10⁻⁸.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationPartial
Human understandingFull

A counterexample to the generalized Viterbo conjecture on the spectral distance in a disc cotangent bundle

Filip Broćić, Baptiste Serraille, Ibrahim Trifa

  • Mathematics
  • Theoretical Physics

The initial negatively curved, Z2-coefficient counterexample to the generalized Lagrangian spectral-distance assertion, as identified in the AI disclosure; exclude the later human extensions.

AI involvementAI-centered
ImportanceModest
Technical verificationReported
Human understandingFull

Positivity and negativity for h-bases for n

Wouter van Doorn, Pietro Monticone, Melvyn B. Nathanson

  • Mathematics

The original comparisons for three-element integer h-bases with exactly h summands: requiring a negative element gives a strictly smaller maximal covered interval for odd h or h=4, and the same maximum for even h other than 4. These are Aristotle’s original proofs, not Nathanson’s replacement arguments.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationReported
Human understandingFull

Explicit positive lower density for the Collatz Conjecture

Kenny Lau

  • Mathematics
  • Theoretical Computer Science

Lau reports an explicit refinement of Mazur’s positive-density Collatz result: writing L=2^(2^(2^87640)), for every X≥2^L at least X/L positive integers n<X reach 1 within floor(10.46 log n) ordinary Collatz steps. The formal development also gives an odd-starting-value version with accelerated-step bounds and target-reaching variants for prescribed positive targets not divisible by three.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFull

The normal order of the number of iterations of Euler’s function

Plasma AI

  • Mathematics

Let f(n) count the iterations of Euler’s totient function needed to reach 1. The manuscript reports a constant α between 1/log 3 and 1/log 2 such that f(n)/log n tends to α on a set of natural density one. It also gives a quantitative mean-square bound, answering the first two parts of Erdős Problem 408.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

ChatGPT-led proof establishes optimality and uniqueness of Romik’s ambidextrous sofa

  • Mathematics
  • Theoretical Computer Science

Every compact connected planar set that can be moved continuously through both a right-turning and left-turning unit-width hallway from the same starting orientation has area at most M=1+4Y²+arctan(Y), where Y>0 solves 4Y³+3Y=1. Equality holds exactly for rotated and translated copies of Romik’s sofa, and the sofa attains M.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean + comparator)
Human understandingFew

Claude-assisted proof gives quantitative stability for optimal qubit cloning

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For any finite-Kraus 1→2 qubit channel whose mean fidelity over the six Pauli eigenstates is at least 5/6−ε, the Kraus operators have total squared Frobenius distance at most 6ε from the span of the two Bužek–Hillery Kraus operators. For every input matrix ρ, the channel output is within 200√ε‖ρ‖F of the Bužek–Hillery channel in Frobenius norm.

AI involvementAI-led
ImportanceModest
Technical verificationFormal (Lean)
Human understandingFew

A Central Limit Theorem for High-Temperature SYK Model

Yoshinori Kamijima, Shuta Nakajima, Ryoichiro Noda, Kôhei Sasaya

  • Mathematics
  • Theoretical Physics

Kamijima, Nakajima, Noda and Sasaya report a nondegenerate Gaussian limit for N^(q/2−1)(log Z_N−E log Z_N) in the Gaussian q-body SYK model, for fixed even q≥4 and sufficiently small nonzero inverse temperature.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

AI-assisted Lean proof resolves the factorial divisor-sum irrationality problem for all exponents

  • Mathematics
  • Theoretical Computer Science

For every natural k≥0, the convergent series Σₙ≥1 σₖ(n)/n! is irrational, where σₖ(n) is the sum of kth powers of positive divisors of n. The k>0 cases answer Erdős Problem 252; the Lean theorem also handles k=0.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

GPT-assisted proof establishes the full last-gradient curve for smooth convex descent

  • Mathematics
  • Theoretical Computer Science

For every N≥0 and h≥0, constant-step gradient descent xᵢ₊₁=xᵢ−(h/L)∇f(xᵢ) on a differentiable convex function with L-Lipschitz gradient in a real Hilbert space satisfies ‖∇f(xN)‖≤L‖x0−x*‖max{1/(1+Nh),|1−h|^N}. One-dimensional Huber and quadratic examples attain the two branches. The compared Lean contract also classifies sampled equality trajectories and proves the same sharp factor for the terminal residual of relaxed proximal minimization with actual minimizers.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

A quasi-Riemann hypothesis for Dedekind zeta functions

Dean Cureton, Jesse Thorner

  • Mathematics
  • Theoretical Computer Science

Cureton and Thorner’s version-two manuscript reports that for every number field K, its Dedekind zeta function has no zeros in Re(s)>31/32, excluding its pole at s=1. A corollary gives the same half-plane for finite-order Hecke L-functions wherever they are holomorphic, including s=1 for nonprincipal characters.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean)
Human understandingFew

Claude-assisted search finds the smallest tournament with capacity above its transitive number

  • Mathematics
  • Theoretical Computer Science

The specified seven-vertex induced Paley tournament has transitive number 4 and capacity between √18 and 5; every tournament on at most six vertices has capacity equal to its transitive number, so seven is minimal.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean + comparator)
Human understandingFew

Frankl-complete sunflowers and extremal families of four-sets

Mingchang Liu

  • Mathematics

Liu’s first release gives FC(4,9)=16 and a co-developed 3:1 core-petal charging argument yielding the nine-petal threshold and FC(4,n)=Θ(n²). A later version extends the asymptotic to FC(k,n)=Θₖ(n^(k−2)) for each fixed k≥2; that general proof is separate from the model-attributed first-release work.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

The in-tree of an n-cycle under cyclic conjugation: depth pᵉ⁻¹ for odd p, and 2ᵉ⁻¹−1 for p=2

Frédéric Lefebvre-Naré

  • Mathematics

For D(q)=q s q⁻¹ on n-cycles, Lefebvre-Naré’s October note claims depth p^(e−1) for odd prime powers p^e, depth 2^(e−1)−1 for powers of two, and the maximum prime-power depth for general n. The shorter proof uses a defect commutator and triangular Sylow-group filtration.

AI involvementAI-led
ImportanceModest
Technical verificationPartial
Human understandingFew

GPT-assisted proof finds exact robust gaps for finite chiral chains

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For a finite open chain with positive strong-bond lower bounds βᵢ, boundary parameter s≥0 and weak-to-strong ratios qᵢ∈[ℓᵢ,ρᵢ], the minimum scalar squared gap minq λmin(DβQ(q)ᵀQ(q)Dβ) is positive and equals the minimum over the ordered one-switch endpoint vectors q⁽ᵏ⁾, with lower endpoints before k and upper endpoints after k. The same value sharply bounds the squared nonzero gap for arbitrary finite-dimensional complex matrix channels satisfying the stated singular-value constraints.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Brill--Noether existence and rank for graphs through genus six

Nathan Pflueger

  • Mathematics

For every connected finite loopless multigraph of genus g≤6 and integers d,r≥0 with ρ=g−(r+1)(g−d+r)≥0, a degree-d divisor of Baker–Norine rank at least r exists. The new case is genus six, proved by descending special divisors from odd subdivisions to the original graph.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean 4/Mathlib)
Human understandingFew

GPT-assisted Lean proof settles the planar case of a dual-midpoint conjecture

  • Mathematics
  • Theoretical Computer Science

On a connected open subset of the Euclidean plane, if the canonical local geodesic midpoint maps of the dual torsion-free connections D+KC and D−KC are complementary near every diagonal point, then their totally symmetric cubic tensor field C is constant. The formal proof constructs the short geodesic branches and also proves a differential obstruction criterion and a five-direction test.

AI involvementAI-led
ImportanceNotable
Human understandingFew

GPT-assisted proof finds the sharp robustness radius for a four-row permanent inequality

  • Mathematics
  • Theoretical Computer Science

For a balanced probability law ν on the 24 permutations of four labels at total-variation distance at most r≤1/24 from uniform, and nonnegative rows fᵢ normalized by the uniform mean of fᵢ² equal to one, 1−Eν∏ᵢfᵢ(σ(i))≥(1−24r)/9 ΣᵢVar(fᵢ). The coefficient is sharp at each radius and 1/24 is the exact admissible radius.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Proudfoot–Perrault reports: a laser-method exponent below 2.3710449

  • Mathematics
  • Theoretical Computer Science

Proudfoot and Perrault report that the exact tensor-rank matrix-multiplication exponent is strictly below 2.3710449 over every commutative ring. Their formal theorem uses the exact rational bound 23710449/10000000; the accompanying paper rounds this to 2.371045.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

Nonquadraticity of Lambert Series Values

Costa Lambrinoudis

  • Mathematics
  • Theoretical Computer Science

For F(q)=Σₙ≥₁1/(qⁿ−1), a degree criterion excludes polynomial relations of degree at most e over Q(q) when eλ log M(q)<log q, where λ=(42963+428220/π²)/202140. It gives nonquadraticity for every integer q≥2 and irrationality for rational q=a/b>1 when log b/log a<0.5728; a certified refinement reaches 0.573290.

AI-led proof gives a linear Betti bound for four-generated semigroups

  • Mathematics
  • Theoretical Computer Science

For a minimally four-generated numerical semigroup S=⟨n₀<n₁<n₂<n₃⟩, the minimal number of generators of its toric ideal is at most 2W over any field, where W=(n₃−n₀)/gcd(n₁−n₀,n₂−n₀,n₃−n₀). Consequently it is at most twice the ordinary width n₃−n₀.

AI involvementAI-led
ImportanceNotable
Human understandingFew

AI agents prove six concyclic points meet a Euclidean Ramsey tensor criterion

  • Mathematics
  • Theoretical Computer Science

Every set of six distinct points on a circle satisfies the FieldCriterion tensor condition of the OpenAI Euclidean Ramsey classification. More generally, every spherical finite set of at most m+4 points, where m is its affine-span dimension, satisfies that criterion in its own coordinate field; every spherical set with algebraic coordinates also satisfies it.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

Navot: AI-assisted extensions of the 9/4 matrix multiplication bound to fields and rings

  • Mathematics
  • Theoretical Computer Science

Navot’s AI-assisted Lean developments extend OpenAI’s matrix multiplication exponent bound ω ≤ 9/4 from complex numbers to every field, including finite fields and positive characteristic, and subsequently to every associative unital ring, including noncommutative rings. The ring development establishes exact integer coefficient tensor-rank identities and, for every ε > 0, correct division-free arithmetic programs with cost at most C n^(9/4 + ε) at every positive matrix size, for one C independent of n. Palomar registered the field development on 8 October and the ring development on 11 October 2026.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

Uniform real Lipschitz surfaces on the triangular lattice

OpenAI

  • Mathematics
  • Theoretical Physics

For uniform real nearest-neighbor Lipschitz heights on triangular-lattice approximations of smooth simply connected domains, the October 6 manuscript claims a centered and normalized Dirichlet Gaussian free-field limit. With a tuned opposite two-arc boundary amplitude, the zero-height interface converges to chordal SLE₄ in uniform curve distance.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

The p-adic section conjecture

OpenAI

  • Mathematics

For every prime p, finite extension k/Q_p and smooth proper geometrically connected curve X/k of genus at least two, the manuscript claims a bijection between X(k) and sections of the full arithmetic étale fundamental-group sequence modulo conjugation by the full geometric group.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a solenoidal-force Navier–Stokes halting detector

  • Mathematics
  • Theoretical Physics

For every deterministic one-tape machine and finite input, an effective smooth divergence-free mean-zero force on the flat three-torus gives a unique classical forced Navier–Stokes flow from rest, with zero pressure, whose fixed material particle enters one fixed open strip exactly when the machine halts. The force and velocity have bounded mixed derivatives and become one-periodic after initialization.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports sharp one-third stability of Brenier maps

  • Mathematics

For the uniform measure on any compact convex body with nonempty interior in ℝᵈ, d≥2, and targets supported in one fixed compact set, the L² distance between their quadratic optimal maps is at most C times the one-third power of the targets’ 2-Wasserstein distance. No larger uniform exponent works in general, even for three-atom targets on a fixed cube.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports nonattainment of the three-marginal Coulomb Monge minimum

  • Mathematics

There is a smooth compactly supported probability density on ℝ³, with smooth compactly supported square root, whose finite three-marginal Coulomb Kantorovich minimum is attained by a coupling but by no pair of measure-preserving Borel maps; nevertheless the Monge and Kantorovich infima are equal.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports an independent proof for reflexive nonexpansive fixed points

  • Mathematics

OpenAI gives an adaptive-anchor and weighted-tree proof that every nonexpansive selfmap of a nonempty closed bounded convex subset of a real reflexive Banach space has a fixed point in the given norm; the manuscript also derives the commuting-family retract corollary from Bruck’s theorem.

AI involvementAI-led
ImportanceModest
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports universal critical SK quench autocorrelation limits

  • Mathematics

For zero-field Sherrington–Kirkpatrick heat-bath dynamics at β=1, stationary and fair-spin quench autocorrelations converge jointly as random functions after multiplying correlations by n^{1/3} and scaling waiting times and lags by n^{2/3}; Gaussian and Rademacher couplings have the same limiting law, and the quench limit relaxes to the stationary limit as waiting time grows.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports a conditional quenched near-critical Voronoi limit

  • Mathematics

Assuming the scalar, annealed Cardy crossing formula for fair-colored critical Poisson–Voronoi percolation on every bounded Jordan quadrilateral, the jointly coupled near-critical crossing thresholds of rational polygonal quadrilaterals, normalized by each model’s expected number of unit-square color-pivotal sites, have conditional laws that converge in probability over Voronoi tessellations to the triangular-lattice limit.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports the all-density Ising perceptron free energy

  • Mathematics

For the Ising perceptron with independent Gaussian patterns, every bounded Borel real log-potential and each fixed positive temperature and pattern density have a limiting normalized free energy given by the manuscript’s explicit overlap-path variational formula, with convergence in expectation and probability.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

OpenAI reports quenched SLE₃ convergence for weak random-bond Ising interfaces

  • Mathematics

For planar Ising bonds J_e=1+εξ_e with any fixed bounded, nondegenerate, mean-zero iid disorder law and sufficiently small ε>0, the full oriented interface law at the spontaneous-magnetization threshold converges in probability over environments to chordal SLE₃ in deterministic Jordan-domain approximations.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a canonical massive continuum limit of the two-dimensional O(3) model

  • Mathematics

For the nearest-neighbor two-dimensional O(3) spin model with no external field or topological term, normalization by susceptibility and second-moment correlation length yields a unique continuum Schwinger hierarchy as the bare inverse coupling tends to infinity through all positive real values; the reconstructed field theory has a positive mass gap and a nonzero connected four-point function.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports Liouville Brownian limits of walks on critical FK–Ising maps

  • Mathematics

For critical spherical FK–Ising planar maps with n edges, stationary random walks using every incident half-edge converge jointly with the rescaled map metric and degree measure to Liouville Brownian motion on the unit-area √3-quantum sphere when time is accelerated by exactly n; the convergence holds for any fixed finite number of conditionally independent walks.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports joint metric and connection recovery from one boundary patch

  • Mathematics

For a compact connected smooth manifold of dimension at least three with smooth boundary and any nonempty open boundary patch, equality of zero-frequency same-patch boundary measurements for two smooth metrics and smooth unitary connections on the trivial rank-two Hermitian bundle implies they differ only by a patch-fixing diffeomorphism and a patch-identity unitary gauge.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports Boltzmann nonuniqueness with exact local conservation

  • Mathematics
  • Theoretical Physics

The three-dimensional periodic hard-sphere Boltzmann equation has one nonnegative initial density with bounded velocity support and finite mass, energy and absolute entropy that yields two distinct global entropy solutions, both satisfying exact local conservation of mass, momentum and kinetic energy and strong L¹ continuity.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports Gigli’s Alexandrov curvature characterization

  • Mathematics

For every integer n≥2 and real κ, a complete separable metric space is an n-dimensional Alexandrov space of curvature at least κ exactly when its Hausdorff-measure metric space has full support, is RCD((n−1)κ,n), and has Gigli distributional sectional curvature at least κ on the original global test classes.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports thermodynamic-first QAOA optimality for the SK model

  • Mathematics
  • Theoretical Physics

For the zero-field Gaussian Sherrington–Kirkpatrick model, as system size tends to infinity first and QAOA depth then grows, the best limiting expected energy per spin approaches the Parisi ground-state value. For every accuracy, one finite depth and deterministic angles work independently of system size and disorder.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports entanglement with zero distillable secret key

  • Mathematics
  • Theoretical Physics

An explicit entangled state on C¹⁰⊗C¹⁰ has zero distillable secret key against an eavesdropper holding a purification and the full public transcript, for completed local-instrument protocols with joint processing of copies and unlimited two-way public communication. Every admitted bit-output protocol remains at trace distance at least 1/10 from an ideal secret bit.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a uniform excess-charge bound for Coulomb molecules

  • Mathematics
  • Theoretical Physics

For any fixed nuclei at distinct positions with real charges at least one, total nuclear charge Z and M nuclei, the full two-spin nonrelativistic Coulomb Hamiltonian can strictly bind no more than Z+CM electrons, where C is universal and independent of the nuclear positions and charges.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports Bloch’s law for finite-range Heisenberg ferromagnets

  • Mathematics
  • Theoretical Physics

For every fixed positive half-integer spin and every nonnegative symmetric finite-range interaction generating ℤ³, the spontaneous magnetization deficit of the three-dimensional quantum Heisenberg ferromagnet is asymptotic to ζ(3/2)T^{3/2}/(8π^{3/2}√det D), where D is the positive quadratic one-magnon dispersion matrix.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports disorder stability of the spherical Laughlin gap

  • Mathematics
  • Theoretical Physics

For the fermionic V₁ Laughlin Hamiltonian at flux q=3(N−1), sufficiently weak projected bounded real scalar one-body potentials leave a unique ground state and a spectral gap bounded below uniformly for all sufficiently large N and all potential profiles of sup norm at most one.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports positive-temperature condensation for dilute hard spheres

  • Mathematics
  • Theoretical Physics

For each hard-sphere radius a>0 and every sufficiently small fixed density ρ>0, there is a temperature T(a,ρ)>0 independent of volume at which the exact three-dimensional canonical Gibbs state has a positive condensate fraction in the constant orbital as N/L³ tends to ρ.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports all equality cases for a sharp matrix Lieb–Thirring bound

  • Mathematics

For every finite matrix size and 1/2<γ<3/2, a measurable positive semidefinite Hermitian matrix potential saturates the sharp one-dimensional Lieb–Thirring inequality exactly when one constant unitary basis makes it a direct sum of independent one-bound-state scalar sech² profiles and zero channels.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports a solution to Howie’s nonsingular-systems conjecture

  • Mathematics

For any group G and any finite system of m equations in n unknowns, with 1≤m≤n, if the exponent-sum matrix has row rank m over the rationals, then G embeds in the group formed by adjoining the unknowns and imposing the equations. Thus the system has a simultaneous solution in an overgroup.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a regular-lifespan Boltzmann–Grad limit for stable radial potentials

  • Mathematics
  • Theoretical Physics

For stable finite-range radial pair potentials in the stated regularity class, grand-canonical Newtonian gases started from spatially summable C¹ data with Gaussian density and spatial-gradient bounds have every fixed scaled factorial marginal converge in full phase-space L¹, uniformly on any finite interval where the corresponding nonlinear Boltzmann equation has a nonnegative classical solution with the required Gaussian bound.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports filtered-product nullity from complex cobordism

  • Mathematics

For a finite pointed CW pair G ⊂ E and based map v:E→Sᵈ whose restriction to G becomes null after passage to complex cobordism, the restriction of a sufficiently large smash power v∧ℓ to the union where any fixed positive fraction of factors lie in G is stably null. The paper also derives a boundary-preserving cube compression theorem from a torus-slot bundle embedding of rank strictly below twice the source rank.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports arithmetic classification of twisted lattice-factor correspondences

  • Mathematics

For an ICC group commensurable with a lattice in a finite product of the specified property-(T) algebraic groups over characteristic-zero local fields, any countably infinite ICC comparison group, and arbitrary normalized scalar cocycles at both ends, every bifinite correspondence between their twisted group factors is a closed bimodule summand of a finite sum of models from finite-index subgroup isomorphisms and finite-dimensional projective representations of the matched cocycle ratio. Conversely, every such model, sum, and closed summand is bifinite.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports local conformal symmetry from scale symmetry in a four-dimensional QFT class

  • Mathematics
  • Theoretical Physics

Under Assumption 2.1 for a four-dimensional local, unitary, positive-energy vacuum QFT—with a physical local stress tensor and dilatation current, discrete bounded-below finite-multiplicity scaling spectrum, finite scaling support of each original field, and compatible affiliated fields in a causally commuting bounded local net—the stress tensor admits a physical dimension-two scalar improvement that is symmetric, conserved, and traceless without changing translation or Lorentz charges. Its conformal currents obey unbroken local Ward identities on the physical field algebra.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports generic weak future inextendibility near rotating Kerr

  • Mathematics
  • Theoretical Physics

For each fixed subextremal rotating Kerr bridge (M > 0 and 0 < |a| < M), smooth vacuum data in a sufficiently small weighted tenth-order neighborhood whose full maximal globally hyperbolic development admits a future continuous, nondegenerate Lorentzian extension with locally square-integrable weak connection form a meagre set. Hence inextendible data are residual in the weighted smooth topology.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports an FPRAS for common integer polymatroid bases

  • Mathematics
  • Theoretical Computer Science

With exact rank-value oracles, the number of common integer bases of two equal-rank polymatroids can be approximated by an FPRAS whose oracle calls and non-oracle bit work are polynomial on every run in ground-set size, binary input length, inverse relative error, and log inverse failure probability.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports uniqueness of first-time surface mean-curvature tangents

  • Mathematics

At every singular point of the first singular time for the mean-curvature flow of a smooth compact connected embedded surface in R³, the full fixed-center rescaling family converges locally smoothly on compact negative-time intervals to one multiplicity-one homothetic self-shrinker in the original ambient coordinates.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports approximation paving for every maximal abelian subalgebra

  • Mathematics

For every maximal abelian subalgebra of a complex von Neumann algebra, every self-adjoint operator is a strong-operator limit of self-adjoint approximants of norm at most three times its norm, each admitting norm paving to relative error ε with at most Cε⁻⁶ projections for a universal C.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a Coulomb density without Kohn–Sham ensemble representation

  • Mathematics
  • Theoretical Physics

A finite three-electron molecule with two equal positive integer nuclear charges has a normalizable interacting Coulomb ground state whose spin-summed density is not the density of any noninteracting ground-state ensemble for a single real spin-independent local potential in L^(3/2)(R³)+L∞(R³).

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports orbifold Iitaka subadditivity

  • Mathematics

For a surjective connected-fiber holomorphic map from a smooth compact connected Fujiki-class-C manifold with rational simple-normal-crossing boundary to a normal compact irreducible complex space, the total log Kodaira dimension is at least the sum of the very general fiber log Kodaira dimension and Campana’s orbifold-base Kodaira dimension, including boundary coefficient one.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports integral density on curve character varieties

  • Mathematics

For every smooth connected complex algebraic curve, every rank r, and prescribed quasi-unipotent boundary conjugacy classes, including nonsemisimple classes, one finite extension of the number field generated by their eigenvalues makes ambient integral points Zariski dense in the exact-boundary locus of the SL_r character variety and in every irreducible component.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a joint Ashkin–Teller height-and-current limit

  • Mathematics
  • Theoretical Physics

At every fixed point on the critical Ashkin–Teller line, including the four-state Potts endpoint, the height and both current-cluster collections converge jointly in bounded Jordan domains: the height to a Gaussian free field and the clusters to canonical recursive two-valued local sets of that field, including all nesting depths and the wired boundary cluster.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports Gaussian free-field limits for weighted Lipschitz heights

  • Mathematics
  • Theoretical Physics

For each fixed edge-change weight x∈[1/√2,1], zero-boundary integer Lipschitz heights on triangular-lattice inside approximations of a bounded C² Jordan domain converge, after division by a positive x-dependent constant, to the zero-Dirichlet Gaussian free field as a random distribution.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports Gaussian free-field limits across the balanced six-vertex range

  • Mathematics
  • Theoretical Physics

For the balanced-torus plane state of the square-lattice six-vertex model with a=b=1 and 0<c≤2, its unit-jump height field converges modulo constants to a Gaussian free field whose squared multiplier is 1/arcsin(c/2) under the Green-kernel normalization −(2π)^(-1)log|x−y|.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports fixed-energy GOE statistics for weakly disordered regular graphs

  • Mathematics

For every fixed degree d≥3, sufficiently weak fixed positive uniform diagonal disorder on a uniformly random simple d-regular graph yields the GOE bulk microscopic eigenvalue point process at each fixed energy separated from the clean spectral edges, after rescaling by the corresponding tree density of states.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports uniqueness in Bogomolov–Pop reconstruction

  • Mathematics

For function fields of transcendence degree at least two over algebraically closed fields of characteristic different from a fixed prime ell, two perfect-closure field isomorphisms preserving the constant fields that induce the same pro-ell abelian-by-central Galois data up to ell-adic unit scaling agree up to Frobenius.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports a two-dimensional entanglement area law

  • Mathematics
  • Theoretical Physics

For finite-range bounded interactions of fixed local dimension on any finite induced subgraph of the square lattice, a unique ground state separated by a uniform global spectral gap has entanglement entropy at most C times the crossing-edge boundary size for every set of sites.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a group without fixed price

  • Mathematics

A finitely generated group has two essentially free probability-measure-preserving actions with different costs: its Bernoulli action has cost strictly greater than one, while a suitable finite-height extension has cost closer to one.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports the lex-plus-powers Betti theorem in characteristic zero

  • Mathematics

Over every characteristic-zero field, for a homogeneous ideal containing a regular sequence of any positive length with degrees at least two, a corresponding lex-plus-powers ideal has the same Hilbert function and at least as large a graded Betti number in every homological and internal degree.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports pointwise multiple ergodic convergence for mixing transformations

  • Mathematics

For every invertible mixing probability-preserving transformation and every fixed finite tuple of bounded measurable functions, consecutive multiple ergodic averages converge almost everywhere, along all positive averaging lengths, to the product of the functions’ integrals.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports global quantum geometric Langlands at irrational level

  • Mathematics

For every connected simple complex algebraic group G, smooth projective connected complex curve X, and complex level c outside Q, the full twisted D-module categories on Bun_G(X) and Bun_G-dual(X) are equivalent at dual shifted levels c and -1/(r c), retaining the given global forms and all connected components.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports numerical semiampleness for nef Kähler adjoint classes

  • Mathematics

On a smooth connected compact Kähler manifold, if B is an effective rational simple-normal-crossing divisor with coefficients below one, M is a nef rational line bundle, K_X+B is pseudo-effective, and K_X+B+M is nef, then its real Bott–Chern first Chern class is represented by a semiample rational line bundle.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports an étale cover with prescribed exterior and disk degrees

  • Mathematics

For a smooth proper hyperbolic curve over an algebraic closure of a p-adic local field and finitely many disjoint sufficiently small open disks, there is one connected finite étale cover with a component mapping isomorphically onto the entire disk exterior and every component above each disk having degree divisible by p.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports conditional global Arthur enhancements at full finite level

  • Mathematics

Assuming the stated finite-level Ramanujan–Arthur decomposition, every occurring cuspidal excursion parameter for a split connected semisimple group over a global function field at a specified full finite level has one algebraic SL2 and commuting Weil centralizer map recovering the parameter on the entire Weil group, including inertia.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports pro-modularity at the prime two

  • Mathematics

Every continuous odd absolutely irreducible two-dimensional 2-adic representation of the absolute Galois group of Q that ramifies at only finitely many finite primes occurs in the full completed varying-weight Hecke algebra at some odd tame level, possibly with auxiliary primes.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports a pinched Kähler threefold without bounded coordinates

  • Mathematics

There is a contractible complex three-dimensional domain with a complete Kähler metric whose real sectional curvatures lie between two negative constants, yet it has no bounded holomorphic map to complex three-space with nowhere-vanishing Jacobian and is not biholomorphic to a bounded domain.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports a reflexive midpoint-convexity counterexample

  • Mathematics

A separable reflexive real Banach space has an asymptotically midpoint uniformly convex given norm with averaged midpoint modulus at least √(1+t²/12)−1 for t>0, but admits no equivalent asymptotically uniformly convex norm; depth-k countably branching diamonds have distortion at least √(1+k/12) in it.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports invariant-projection counterexamples for irrational rotations

  • Mathematics

For every irrational rotation angle θ in (0,1), there is a continuous nonnegative circle weight f with one zero and integral log f equal to negative infinity such that the weighted rotation U f(V) has no nontrivial invariant projection in its associated hyperfinite type II₁ factor.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports an all-temperature pressure formula for orthogonally invariant Ising models

  • Mathematics

For Ising interactions with Haar orthogonal eigenvectors, empirical eigenvalues converging to a compactly supported law, and extreme eigenvalues converging to the support edges, the normalized pressure converges in expectation and almost surely at every fixed positive temperature to the paper’s overlap-hierarchy variational functional.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports a continuum temperature singularity for a radial pair potential

  • Mathematics
  • Theoretical Physics

One stable radial pair potential in three dimensions with a divergent repulsive core and integrable o(r⁻³) tail has canonical free energy at all positive temperatures and densities, with a strict finite jump in its inverse-temperature derivative at one common positive temperature over a nonempty open interval of densities.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports a tensor classification of finite Euclidean Ramsey sets

  • Mathematics

A finite Euclidean configuration is Ramsey for monochromatic congruent copies at its original scale exactly when its coordinate field admits the paper’s rational-tensor matrix satisfying the pointwise evaluation identities and identity-matrix product identities; the criterion covers every affine dimension.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports exact-budget recovery for metric k-median

  • Mathematics
  • Theoretical Computer Science

On positive integral metrics of polynomially bounded diameter, a supplied anchor with distinct accurate proxies for all but logarithmically many comparison clusters and sufficiently small total proxy error supports a polynomial-time k-median recovery algorithm opening at most k facilities and attaining a (1+2/e+ε) factor against those comparison centers with any requested confidence.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports termination of generalized log canonical Kähler fourfold flips

  • Mathematics

Every sequence of generalized log canonical flips on a normal globally Weil Q-factorial compact Kähler fourfold terminates when each flip is a projective small diagram with the stated opposite ample signs, rational boundary, and one fixed nef Q-Cartier b-divisor on a projective modification.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

OpenAI reports equivariant Hikita for arbitrary quivers

  • Mathematics

For every finite quiver, including loops and multiple arrows, and every regular stability character with stable semistable locus and free gauge action, the equivariant cohomology algebra of its Nakajima variety is canonically isomorphic to the coordinate ring of the scheme-theoretic stability-cocharacter fixed locus on the flavor-deformed Coulomb branch.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

OpenAI reports a zero dimension exponent for well-conditioned log-concave sampling queries

  • Mathematics

For C² potentials with known minimizer at zero and I≤∇²V≤2I, exact value-and-gradient oracle sampling to total-variation error at most 1/10 uses at most Cε d^ε queries for every ε>0, while every algorithm needs at least c log d queries eventually.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports a doubling Hilbert set with no finite-dimensional bi-Lipschitz embedding

  • Mathematics

A subset of real Hilbert space has doubling constant at most 76,800 yet admits no bi-Lipschitz embedding into any finite-dimensional Euclidean space at any finite distortion. Moreover, every infinite-dimensional real Banach space contains a compact doubling subset with the analogous obstruction and a universal doubling bound.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports the Euclidean Steinitz–Bergström bound

  • Mathematics

Every prescribed-order finite sequence of vectors in the Euclidean unit ball of R^d has one choice of signs whose every signed prefix has norm at most C sqrt(d), with C independent of sequence length. Consequently every indexed zero-sum family admits an ordering with every unsigned prefix bounded by C sqrt(d).

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

OpenAI reports the Margulis–Platonov conjecture over global function fields

  • Mathematics

For an absolutely almost simple simply connected algebraic group over any global function field, including characteristic two, every noncentral abstract normal subgroup of its rational points is the inverse image of an open normal subgroup of the finite product of its anisotropic local groups.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

A counterexample to Hadwiger's conjecture

OpenAI

  • Mathematics

The manuscript reports arbitrarily large finite simple graphs with independence number at most two and chromatic number greater than their largest clique-minor order. It also claims violation of the ordinary fractional-coloring weakening.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Milne's rationality conjecture for abelian varieties

OpenAI

  • Mathematics

OpenAI reports that rational Hodge classes on abelian varieties with good reduction pair rationally with complementary divisor products on the reduction, compatibly across prime-to-p and crystalline realizations, including residue characteristic two.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The Deligne-Drinfeld conjecture

OpenAI

  • Mathematics

OpenAI reports that the rational Grothendieck–Teichmüller Lie algebra is free on one generator in each odd weight starting at three, with the corresponding weight-completed isomorphism.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Fujita's freeness conjecture

OpenAI

  • Mathematics

OpenAI reports that K_X+mL is globally generated on every smooth complex projective n-dimensional variety for every ample line bundle L and every integer m at least n+1.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Nagata's conjecture for plane curves

OpenAI

  • Mathematics

OpenAI reports that, at r at least ten very general complex plane points, the total multiplicity of every nonzero effective degree-d curve is strictly less than d times the square root of r.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The Gaussian propeller bound in every dimension

OpenAI

  • Mathematics

OpenAI reports that the sum of squared Gaussian first-moment lengths of the cells of every finite measurable Euclidean partition is at most 9/(8π), with the three-sector planar construction attaining the bound in the stated dimensions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The Unique Games Theorem

OpenAI

  • Mathematics
  • Theoretical Computer Science

OpenAI reports a deterministic polynomial-time reduction from 3SAT to fixed-alphabet Unique Games with completeness at least 1−ε and soundness at most δ for each fixed ε,δ in (0,1/2).

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Deterministic Polynomial Factorization over Prime Fields

OpenAI

  • Mathematics
  • Theoretical Computer Science

OpenAI reports uniform deterministic complete factorization of a densely represented polynomial over a binary-specified prime field in bit complexity polynomial in (degree+1) log p, including irreducible factors and multiplicities.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

A counterexample to Sidorenko's conjecture

OpenAI

  • Mathematics

OpenAI reports a bipartite graph with 35 vertices and 66 edges whose homomorphism density in a finite simple host is below the predicted edge-density bound; it also reports failure of the associated forcing assertion.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Rigidity of the Turing degrees

OpenAI

  • Mathematics
  • Theoretical Computer Science

OpenAI reports that every order automorphism of the full Turing-degree partial order is the identity in ZFC, without a definability or regularity restriction on the automorphism.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The Singer conjecture in dimension four

OpenAI

  • Mathematics

OpenAI reports vanishing of universal-cover L²-Betti numbers outside degree two for closed connected aspherical topological four-manifolds, and more generally finite aspherical integral Poincaré complexes of formal dimension four.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Carlet's cyclic-additive conjecture for the Kasami monomials

Gábor P. Nagy, Douglas S. McNeil, Attila Vajda

  • Mathematics

Let K be a field of characteristic two with |K|=2ⁿ, gcd(k,n)=1 and dₖ=4ᵏ−2ᵏ+1. If Δ is the image of b↦(b+1)^dₖ+b^dₖ+1, then for distinct nonzero v₁,v₂∈K exactly 2^(2n−3) triples (x,y,z)∈Δ³ satisfy v₁x+v₂y+(v₁+v₂)z=0.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

The spacetime Penrose inequality with charge and original-data rigidity

OpenAI

  • Mathematics
  • Theoretical Physics

For the manuscript’s smooth one-ended three-dimensional Einstein–Maxwell exterior data satisfying charged dominant energy and weak future trapping assumptions, invariant ADM mass obeys m≥Q and √(A/4π)≤m+√(m²−Q²), where A is minimum enclosing area in the original metric.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Strong Ulam Stability Characterizes Amenability

OpenAI

  • Mathematics

For every countable discrete group, the manuscript claims amenability is equivalent to strong Ulam stability: sufficiently accurate unitary almost representations on any complex Hilbert space are uniformly close in operator norm to genuine representations on the same space. It gives a fixed-distance obstruction for nonamenable groups.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Essential Singularity of the Correlation Length in the Planar XY Model

OpenAI

  • Mathematics
  • Theoretical Physics

For the nearest-neighbor cosine XY model on the square lattice, the manuscript claims sqrt(βc−β) log ξ(β) tends to a finite positive model-specific constant as β rises to βc from below. Here ξ is defined from the inverse axis mass after the infinite-volume and separation limits in that order.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Quantum Depletion and Momentum Distribution in the Dilute Hard-Sphere Bose Gas

OpenAI

  • Mathematics
  • Theoretical Physics

For exact pure and mixed ground states of a three-dimensional hard-sphere Bose gas, the nonzero-momentum occupation measure approaches the scaled Bogoliubov profile: first take the thermodynamic limit at fixed density, then the dilute limit. Its mass gives depletion 8/(3√π)√(ρa³)+o(√(ρa³)), uniformly over thermodynamic accumulation values.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Temperedness at ramified places for globally generic exceptional groups

OpenAI

  • Mathematics

For globally generic cuspidal complex automorphic representations of split connected adjoint absolutely simple groups of types G₂, F₄, E₆, E₇ or E₈ over global function fields, every local component is tempered, including arbitrarily ramified places and without a characteristic restriction.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Close Separable C*-Algebras Without Spatial Conjugacy

OpenAI

  • Mathematics

For every epsilon>0, the OpenAI-authored manuscript constructs unital norm-separable C*-algebras on a common separable Hilbert space, with common identity and Kadison-Kastler distance below epsilon, which share their von Neumann closure but are not conjugate by any unitary. It does not assert abstract nonisomorphism or a von Neumann counterexample.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Near Inclusions of von Neumann Algebras Without Small Spatial Embeddings

OpenAI

  • Mathematics

The OpenAI-authored manuscript constructs unital von Neumann algebra pairs on separable Hilbert spaces with one-sided near-inclusion gaps tending to zero, yet every unitary embedding the source into the target stays a fixed distance from the identity. Spatial embeddings do exist; this does not refute two-sided Kadison-Kastler stability.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Truly Subquadratic 3SUM and Truly Subcubic APSP via Triangles in Sparse Lopsided Graphs

Josh Alman, Virginia Vassilevska Williams

  • Mathematics
  • Theoretical Computer Science

Claude’s original sparse-output thin-matrix multiplication algorithm, as presented in Section 2 and related to the certified 3SUM/APSP theorems. Human-optimized numerical parameters, new reductions, data structures and Hinted OMv extensions retain human credit.

AI involvementAI-autonomous
ImportanceMajor
Technical verificationFormal (Lean)
Human understandingFull

AI-developed proof of the Herzog–Schönheim conjecture for solvable groups

  • Mathematics
  • Theoretical Computer Science

In an exact partition of any finite or infinite solvable group into at least two left cosets, two of the corresponding subgroups have the same index. The cardinality form of Erdős Problem 274 follows.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

Optimal packings of one to seven unit squares in a disk

The-Anh Vu-Le

  • Mathematics

For five, six and seven freely rotated, interior-disjoint unit squares in a closed disk, the project reports exact least radii √(5/2), √q* and √13/2, respectively, where q* is its specified quartic root. It classifies all optimal packings up to rotation and relabelling: the five- and six-square optima are unique, while seven squares admit a three-parameter family.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Non-freezing of the proximity digraph and non-pseudostable convergence in heterogeneous Hegselmann--Krause models: Counterexamples to conjectures and results of Mirtabatabaei and Bullo

Peter Hegarty, Damiano Ognissanti, Edvin Wedin

  • Mathematics

Hegarty, Ognissanti and Wedin give convergent heterogeneous Hegselmann–Krause trajectories in the synchronized bounded-confidence and bounded-influence models whose proximity digraphs never stabilize, plus constant-digraph trajectories that fail eventual pseudostability. These counterexamples refute Mirtabatabaei and Bullo Conjectures 2.2 and 2.3 and Theorem 6.4(iv).

AI involvementAI-led
ImportanceNotable
Human understandingFull

Codex reports a sharp two-size Gasoline rounding bound

  • Mathematics
  • Theoretical Computer Science

For positive-integer two-size Gasoline instances with deliveries in {1,K}, the position-first iterative-rounding algorithm under the doubly stochastic assignment-LP interpretation has capacity at most OPT+K−2. For K≥3, the original-index tie rule has matching instances with OPT=K and returned capacity 2K−2, giving worst-case ratio 2−2/K.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

Global Hölder estimates for linearized Monge–Ampère equations in divergence form with applications to dual semigeostrophic equations and periodic homogenization

Guoqing Cui, Chong Gu, Nam Q. Le, Ling Wang, Bin Zhou

  • Mathematics

For n≥3, Proposition 1.2 bounds the spatial L¹ norm of the gradient of the Dirichlet Green’s function for the linearized Monge–Ampère operator by C h^β, uniformly in its pole on a section of height h. The convex C³ potential has Hessian determinant between positive constants; the nested sections satisfy the paper’s compact-containment hypotheses.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Astra-led proof extends a Vietoris obstruction to countable ordinals

  • Mathematics

For every ordinal ω<α<ω₁, Almeida’s compact-range Vietoris power K(α,ord), with its range-plus-finite-coordinate topology, is second-countable and Lindelöf but neither σ-compact nor Menger. The Lean theorem ordinal_full covers this interval by a delayed diagonal.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean 4.19.0)
Human understandingFull

Uniqueness of four-body convex central configurations for all masses

Tejasvi Singh Tomar

  • Mathematics

For four positive masses in the planar Newtonian problem, each cyclic order has exactly one strictly convex central configuration up to similarity. The v2 theorem also gives Q≥K/4 at every convex configuration, nondegenerate local minima and analytic dependence.

Claude-led proof answers Péterfalvi’s p=3 question

  • Mathematics
  • Theoretical Computer Science

For p=3 and every prime q≠2, no finite or infinite group G satisfies Glauberman–Norton Hypothesis (B): it would need an embedding of H=(additive F_{3^q})⋊(norm-one units), a finite abelian subgroup Q of order prime to 3 and y∈Q, with the embedded prime-field line normalizing Q and its y-conjugate normalizing the embedded U.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Full support of the zero-temperature Sherrington-Kirkpatrick order parameter

OpenAI

  • Mathematics
  • Theoretical Physics

The September manuscript proves that every admissible integrable zero-temperature Parisi minimizer for the pure zero-field Sherrington–Kirkpatrick model, with covariance ξ(t)=t²/2, has full relative Stieltjes support on [0,1). It uses a rational coercive identity and strict curvature on constant intervals.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4 (source-reported))
Human understandingFew

The spacetime Penrose inequality and enclosing area

OpenAI

  • Mathematics
  • Theoretical Physics

OpenAI reports the sharp invariant-mass lower bound in terms of minimum enclosing area for one-ended spacetime initial-data exteriors in every dimension n≥3, under the paper’s energy, trapping and decay assumptions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

An explicit exact Hausdorff gauge for SLE

OpenAI

  • Mathematics

For each fixed 0<κ<8, let d=1+κ/8. The September 26 manuscript claims that the explicit small-radius gauge r^d(log log(1/r))^((2−d)/2) gives positive finite Hausdorff measure to every compact positive-time segment of ordinary chordal SLEκ almost surely, with finite expected measure of the full trace inside bounded disks.

AI involvementAI-led
ImportanceMajor
Technical verificationPartial
Human understandingFew

Forbidding r-extension of C₄ — a counterexample of the Anstee–Sali conjecture

Richard P. Anstee, Attila Sali, Pei Wu

  • Mathematics

For each r≥4, the r-extension of the four-cycle has product exponent X(F)=r but admits Ω(n^(r−1/2)) F-free configurations, exceeding the conjectured exponent r−1. The AI-origin contribution is identifying this counterexample family.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

The exact reconstruction threshold for the three-state symmetric channel

OpenAI

  • Mathematics

For the three-state symmetric channel Kλ(i,i)=(1+2λ)/3 and Kλ(i,j)=(1−λ)/3 for i≠j, with −1/2≤λ≤1, reconstruction from depth-ℓ spins and the observed tree on every regular b-ary tree (integer b≥2) and observed Poisson(d) tree (d>1) occurs exactly when dλ²>1 (d=b in the regular case). At admissible equality there is nonreconstruction. Poisson advantage averages over the tree without conditioning on survival.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Variable-dimension Weisfeiler–Leman equivalence on general and subcubic graphs

OpenAI

  • Mathematics
  • Theoretical Computer Science

For joint-update k-dimensional Weisfeiler–Leman equivalence, with k≥2 binary encoded as part of the input and graphs given by explicit adjacency matrices, the test is EXPTIME-complete. The hardness also holds for connected equal-order simple uncolored graphs of maximum degree three.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

Arithmetic Stein-degree bounds for log Calabi–Yau pairs

OpenAI

  • Mathematics

For each dimension d and t>0, the manuscript claims a bound N(d,t) on [k_S:k] for a boundary prime S with coefficient at least t on an ordinary projective log-canonical Q-pair (X,B) over a characteristic-zero field, where X is normal integral, H⁰(X,O_X)=k, B is effective and K_X+B is Q-linearly trivial.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

One-sided negative sectional curvature and the holomorphic Liouville property

OpenAI

  • Mathematics

The OpenAI manuscript claims a domain in some sufficiently large fixed finite complex dimension, diffeomorphic to real Euclidean space, with a complete Kähler metric of real sectional curvature at most −1 and only constant bounded holomorphic functions. The curvature is not bounded below.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Virtual compact specialness of hyperbolic one-relator groups

OpenAI

  • Mathematics

Every word-hyperbolic one-relator group is virtually compact special: a finite-index subgroup is the fundamental group of a finite nonpositively curved cube complex locally isometric to a Salvetti complex. With Kielak–Linton this yields virtual free-by-cyclic structure, allowing an infinite-rank free kernel.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

A logarithmic independence bound for clique-free graphs

OpenAI

  • Mathematics
  • Theoretical Computer Science

Theorem 1.1 claims that for every fixed integer r ≥ 4, each finite K_r-free graph on n vertices with average degree d ≥ 2 has an independent set of size at least c_r n log(d)/d, with c_r > 0 independent of n and d. It removes a log-log denominator in the fixed-clique-size average-degree bound; the paper does not optimize c_r or give an efficient algorithm.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean)
Human understandingFew

Codex project gives exact Thue–Morse matching heights

  • Mathematics

For the Thue–Morse target in Alexander’s binary avoiding population, the project gives an attained exact maximum matching-path length L(v) for every natural starting vertex v, including zero, through a seven-case formula based on the factorization of ⌊v/2⌋+1 and adjacent digit-parity bits. It also proves a finite digit recurrence.

AI involvementAI-centered
ImportanceModest
Human understandingFew

Generalized Star Height at Most Three

OpenAI

  • Mathematics
  • Theoretical Computer Science

Every regular language over a finite alphabet has a generalized regular expression over that same alphabet with at most three nested Kleene stars, allowing union, concatenation and complement.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

An exponential two-way deterministic state lower bound for one-way liveness

OpenAI

  • Mathematics
  • Theoretical Computer Science

For each h≥2, an h+3-state one-way nondeterministic automaton recognizes liveness of products of binary relations, while every equivalent two-way deterministic automaton with s states satisfies 2^⌊(h−2)/31⌋≤4(s+2)². The finite alphabet grows with h.

AI involvementAI-led
ImportanceMajor
Human understandingFew

Deterministic quasipolynomial-time mean-payoff games

OpenAI

  • Mathematics
  • Theoretical Computer Science

For an ordinary mean-payoff game of complete binary input length L with arbitrary signed binary edge weights, the September 25 deterministic manuscript reports a 2^O((log(L+2))²)-bit-operation algorithm for the exact winning set, values and optimal positional strategies.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Short Egyptian fractions

OpenAI

  • Mathematics

The manuscript reports that every a/b with 1 ≤ a < b and sufficiently large b has an expansion with O(log log b) distinct positive unit fractions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFull

Prescribed large-volume charges on threefolds with trivial canonical bundle

OpenAI

  • Mathematics

For each smooth projective complex threefold X, ample integral H and rational B₀, the manuscript gives k≥0 and R*>0 depending only on (X,H,B₀) such that every finite-slope, zero-tilt-slope semistable E satisfies ch₃^(B₀+bH)(E)≤(a²/6+k)H²ch₁^(B₀+bH)(E) for all a>0 and real b. The k correction vanishes uniformly in b for a>R*.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

QMA-hardness of continuum Coulomb energy with unit nuclear charges

OpenAI

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

The electronic Coulomb ground-energy promise problem is QMA-hard for distinct rational positions of unit-charge nuclei, unary electron count and rational thresholds separated by at least an inverse polynomial, with minimization over the full three-dimensional spinful fermionic continuum.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

Elliptic squares and Zilber–Pink for curves in A2

OpenAI

  • Mathematics

Three OpenAI manuscripts jointly claim Zilber–Pink for Hodge-generic curves in the Siegel threefold A₂ over algebraic number fields without a boundary hypothesis. The combined result uses separate finiteness theorems for non-CM elliptic squares, E×CM cases and quaternionic-division points.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Virasoro Constraints for Projective Complete Intersections

OpenAI

  • Mathematics
  • Theoretical Physics

OpenAI claims that every smooth complete intersection in complex projective space satisfies ordinary unreduced descendant Virasoro constraints coefficientwise in every genus and curve class, with arbitrary cohomology insertions including primitive and odd classes.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Hellinger contraction with arbitrary Boolean output bias

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every Boolean function on the uniform discrete cube, including biased outputs, and every product-noise correlation ρ∈[−1,1], the manuscript claims H(Ef)−E H(Tρf)≤1−H(ρ), where H(t)=sqrt(1−t²). Signed coordinates attain equality.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Quantitative lower bounds for trace reconstruction

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every fixed deletion probability q∈(0,1), worst-case reconstruction of deterministic binary strings from independent deletion traces needs n^{Ω(log log n)} samples for any fixed positive success probability, even with unrestricted computation.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

Exact Fourier certificates for complex Hadamard matrices of order six

OpenAI

  • Mathematics

For every order-six complex Hadamard matrix outside Tao’s cubic equivalence class, the normalized character sum vanishes at every permutation of (1,1,1,−1,−1,−1), allowing row and column permutations and arbitrary unit phases.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Minimal models and Mori fibre spaces for generalized log canonical Q-pairs

OpenAI

  • Mathematics

For projective generalized log-canonical Q-pairs over an algebraically closed characteristic-zero field with a fixed nef rational b-divisor, the manuscript claims a birational model that extracts no divisors and preserves discrepancy inequalities, with nef adjoint when pseudo-effective, or a relative-Picard-number-one Mori fibre space otherwise.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Polynomial PEPS approximation of gapped square-grid ground states

OpenAI

  • Mathematics
  • Theoretical Physics

For fixed local dimension q, interaction bound J and global spectral gap Δ>0, the manuscript claims a PEPS with bond dimension at most CLᶜ and normalized vector error at most L⁻¹ for the unique ground state of any on-site plus nearest-neighbour Hamiltonian on an open L×L square grid.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

The Bass trace conjecture and the characteristic-zero Kaplansky idempotent conjecture

OpenAI

  • Mathematics

For every discrete group G, OpenAI claims that the Hattori–Stallings trace of every class in K₀(C G) vanishes on each infinite-order conjugacy class. The claim concerns the complex group ring; an integral identity-only consequence needs a separate known finite-order input.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean)
Human understandingFew

Smooth anisotropic uniqueness in the Calderón problem from one boundary patch

OpenAI

  • Mathematics

For compact connected smooth manifolds of dimension at least three, the manuscript claims that equal density-valued energy Dirichlet-to-Neumann forms measured on any nonempty proper relatively open boundary patch determine two smooth metrics up to a smooth diffeomorphism fixing that patch.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The Poisson-Dirichlet law for prime predecessors

OpenAI

  • Mathematics

For primes p≤x with equal weight, the normalized logarithms of the ordered prime factors of p−1, counted with multiplicity, converge in every fixed finite joint distribution to Poisson–Dirichlet(1). The manuscript also derives a fixed-threshold Dickman law for the largest factor.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

The approximation threshold for metric k-median

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every fixed ε>0, a deterministic polynomial-time algorithm on finite rational metric k-median instances with specified candidate facilities opens at most k facilities and attains factor 1+2/e+ε. Together with known P≠NP hardness, this gives infimum approximation factor 1+2/e.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

Rationality of the Canonical Unramified Arthur Filtration

OpenAI

  • Mathematics

Assuming the four stated characteristic conditions on the split connected semisimple group G over a finite field, every canonical unramified Arthur-filtration subspace on the full space of finitely supported automorphic functions is defined over Q and is independent of ℓ, including noncuspidal functions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The affine Bernstein theorem in dimensions three through nine

OpenAI

  • Mathematics

The OpenAI manuscript claims that, for dimensions 3 through 9, every smooth locally uniformly convex affine-maximal graph complete in its induced Euclidean metric is an elliptic paraboloid. The manuscript also derives the affine-metric-complete immersed-hypersurface corollary. The companion smooth nonquadratic entire example in dimension 10 is a separate contribution and does not supply affine-metric completeness.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Polynomial Hitting Lists for Noncommutative Rational Formulas

OpenAI

  • Mathematics
  • Theoretical Computer Science

For rational-coefficient affine linear matrix pencils, a universal nonadaptive polynomial-size list of rational matrix evaluations detects nonsingularity over the free skew field. The paper also gives a rank-detecting list for rectangular rational pencils, with polynomial construction and output bounds in the dimensions and variable count.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Microscopic jamming in the negative spherical perceptron

OpenAI

  • Mathematics
  • Theoretical Physics

For the spherical perceptron at margin −1 with quadratic penalty, after successive limits in system size, inverse temperature and density approaching the feasibility threshold from above, the gap and force laws have exponents γ=1/(2+θ), with 0.4126930<γ<0.4126934 and 0.4231063<θ<0.4231088. The same limits are claimed for a sufficiently small fixed Gaussian scale mixture.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

A spectral gap throughout the high-temperature Sherrington–Kirkpatrick phase

OpenAI

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For the zero-field Gaussian Sherrington–Kirkpatrick model at every fixed 0<β<1, rate-one-per-site heat-bath dynamics satisfy a dimension-independent Poincaré inequality for all observables with disorder probability tending to one. Uniformly chosen single-site updates have gap at least 1/(Cβ n).

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

The Restricted Geometric Langlands Equivalence in Positive Characteristic

OpenAI

  • Mathematics

For a connected reductive group on a smooth projective connected curve over an algebraically closed field of characteristic p>0, the restricted geometric Langlands functor reaches every spectral component under either finite-field descent or, with p very good and p not dividing the Weyl-group order, an arbitrary algebraically closed base. Both regimes require a nondegenerate invariant Lie-algebra form (also on Levi centers), Chevalley restriction for every Levi, semisimple centralizers that are Levi, and nilpotents lying in parabolic unipotent radicals over every field extension.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

A Fock-space inequality and the Laughlin spectral gap

OpenAI

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For the fermionic Laughlin state at filling 1/3 on the round sphere and flux 3(N−1), the unperturbed V1 Hamiltonian has a uniform gap for all sufficiently large particle numbers N. The released Lean scope gives a 1/25 Fock-space distance estimate and a 1/100 finite spherical gap.

AI involvementAI-led
ImportanceMajor
Human understandingFew

Spontaneous magnetization in the quantum Heisenberg ferromagnet

OpenAI

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For the nearest-neighbor isotropic quantum Heisenberg ferromagnet on ℤᵈ with d≥3 and every positive half-integer spin S, a translation-invariant zero-field KMS equilibrium state exists at sufficiently low positive temperature with magnetization at least S/4.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

Brownian continuum random tree limits of finite Fortuin–Kasteleyn maps above four

OpenAI

  • Mathematics
  • Theoretical Computer Science

For each fixed q>4, critical finite Fortuin–Kasteleyn planar maps with n edges, distances scaled by a q-dependent constant times n⁻¹ᐟ², and normalized degree measure converge in distribution to the Brownian continuum random tree in Gromov–Hausdorff–Prokhorov topology.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4)
Human understandingFew

Triangular minimality for planar Coulomb renormalized energy

OpenAI

  • Mathematics
  • Theoretical Physics

For the manuscript’s unit-background planar Coulomb renormalized-energy class and admissible square-cutoff families, every admissible field has energy at least that of the density-one triangular lattice. A prior equivalence then identifies the linear term of optimal spherical logarithmic energy.

AI involvementAI-led
ImportanceMajor
Technical verificationPartial
Human understandingFew

A counterexample to Ryser's covering conjecture

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every sufficiently large prime s≡2 mod 3 and sufficiently large odd n depending on s, the September 23 manuscript constructs an intersecting (s^n+1)-partite uniform hypergraph with matching number one and covering number s^n+1, exceeding Ryser’s proposed bound s^n.

AI involvementAI-led
ImportanceMajor
Human understandingFew

A counterexample to Ryser's covering conjecture

OpenAI

  • Mathematics

For every sufficiently large prime s≡2 mod 3 and sufficiently large odd n depending on s, with r=s^n+1, the manuscript constructs an r-edge-colored complete graph requiring exactly r monochromatic trees to cover its vertices, even when trees may overlap and singletons count.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

The strong hyperkähler SYZ conjecture

OpenAI

  • Mathematics

Every holomorphic line bundle with nonzero nef isotropic first Chern class on a compact irreducible holomorphic symplectic Kähler manifold is semiample, according to the OpenAI manuscript. The resulting map has a normal projective base of half the complex dimension and Lagrangian fibers.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Translative covering densities of order n log n

OpenAI

  • Mathematics

For every sufficiently large dimension n, an OpenAI manuscript constructs a centrally symmetric convex body whose unrestricted translative covering density exceeds c n log n for an absolute c>0. With Rogers’ O(n log n) upper bound, this fixes the worst-case unrestricted order and rules out a universal O(n) bound.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Maximal Seshadri constants on arbitrary polarized surfaces

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every smooth integral complex projective surface S with ample L, the manuscript claims that all sufficiently large r have maximal multipoint Seshadri constant ε(S,L;p)=√(L²/r) at very general distinct r-tuples, with a nef boundary divisor on the blowup.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Positively curved Einstein four-manifolds

OpenAI

  • Mathematics

The manuscript claims that every connected smooth closed Einstein four-manifold with strictly positive sectional curvature is, up to positive scaling, round S⁴, Fubini–Study CP², or round RP⁴, without assuming orientability.

AI involvementAI-led
ImportanceMajor
Technical verificationPartial
Human understandingFew

Cuntz comparison and Jiang–Su absorption

OpenAI

  • Mathematics

For simple separable unital infinite-dimensional nuclear C*-algebras, the companion manuscripts claim strict comparison implies Jiang–Su absorption, completing the unital Toms–Winter equivalence with finite nuclear dimension. A separate extended-functional theorem also treats nonunital algebras under its own assumptions.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Nonuniqueness for bounded measurable scalar conductivities in three dimensions

OpenAI

  • Mathematics
  • Theoretical Computer Science

The manuscript constructs two distinct uniformly positive bounded measurable scalar conductivities on the ball B(0,3) in R³. Both equal one near the boundary, differ on a set of positive measure and have the same full weak Dirichlet-to-Neumann operator.

AI involvementAI-led
ImportanceMajor
Human understandingFew

Tracial projection methods and uniform property Gamma

OpenAI

  • Mathematics
  • Theoretical Computer Science

For a simple separable unital infinite-dimensional nuclear stably finite C*-algebra with traces, real rank zero of the uniform tracial ultrapower of its uniform tracial completion implies a central projection halving every specified trace on every coefficient. This yields uniform property Gamma for each specified free ultrafilter.

AI involvementAI-led
ImportanceNotable
Human understandingFew

GOE bulk universality for fixed-degree random regular graphs

OpenAI

  • Mathematics
  • Theoretical Physics

For every fixed degree d≥3 and energy strictly inside the Kesten–McKay bulk, the unfolded adjacency-eigenvalue point process of a uniform simple labelled d-regular graph converges at that fixed energy to the GOE bulk process along all admissible graph sizes, without extra conditioning.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Positive scalar curvature forces rational inessentiality

OpenAI

  • Mathematics

Every closed connected oriented smooth manifold of dimension at least two with positive scalar curvature is rationally inessential, without spin or fundamental-group restrictions. The aspherical obstruction also covers nonorientable manifolds through the orientation double cover.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Pure-Point Spectrum for the Two-Dimensional Anderson Model at Every Positive Disorder

OpenAI

  • Mathematics
  • Theoretical Physics

The OpenAI manuscript claims that the nearest-neighbor Anderson operator on ℓ²(Z²), with independent site potentials uniform on [−h,h], has almost surely pure-point spectral type and a complete eigenbasis for every fixed h>0. It also identifies the spectrum as [−4−h,4+h].

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Hard-sphere fluctuations on the regular Boltzmann lifespan

OpenAI

  • Mathematics
  • Theoretical Physics

For the stated deterministic grand-canonical three-dimensional hard-sphere gas with smooth spatially summable Gaussian-bounded initial density and gradient, exactly centered empirical fluctuations converge at every fixed finite set of times and tests to a Gaussian linear fluctuating Boltzmann field on each prescribed finite interval where the classical Boltzmann solution has uniform Gaussian velocity decay.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

A universal group of type F∞

OpenAI

  • Mathematics

OpenAI claims a group of type F∞ containing every finitely presented group. Its finitely generated subgroups, up to isomorphism, are exactly the finitely generated recursively presented groups.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

An infinite finitely presented periodic group

OpenAI

  • Mathematics

The September OpenAI manuscript claims an infinite-dimensional finitely presented nonunital nil associative algebra over F₂ that is Jacobson radical but not nilpotent. Its unitization is finitely presented, algebraic and infinite-dimensional.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

A sharp Fourier certificate for planar circle packing

OpenAI

  • Mathematics
  • Theoretical Computer Science

A real radial Schwartz function on the plane has nonnegative Fourier transform, is nonpositive beyond unit distance, and has normalized origin values making the Cohn–Elkies linear-programming bound attain the triangular-packing density π/(2√3).

AI involvementAI-led
ImportanceNotable
Human understandingFew

Entropy and Face Dimension of the Perfect-Matching Polytope

OpenAI

  • Mathematics
  • Theoretical Computer Science

For a loopless multigraph on 2m vertices with a perfect matching and feasible edge marginals x, the maximum entropy H(x) satisfies F(x)−(2−2/m)B(x)≤H(x)≤F(x), where F and B sum the binary-coordinate entropy terms. Optimizing gives the linear-error estimate proposed by Anari, Oveis Gharan and Vinzant.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Directional transience implies ballisticity

OpenAI

  • Mathematics
  • Theoretical Computer Science

In every dimension d≥2, an IID uniformly elliptic nearest-neighbor random walk environment on Zᵈ with almost-sure transience in a fixed direction has an almost-sure annealed velocity that is deterministic and has positive projection in that direction.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFew

Exponential decay in two-dimensional classical O(n) models

OpenAI

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For every integer n≥3 and finite β>0, spin two-point correlations in finite free-boundary nearest-neighbor square-lattice O(n) models with edge strengths in [0,β] obey a uniform A exp(−m|x−y|) bound. Free-boundary limits inherit the decay.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (Lean 4 / mathlib)
Human understandingFew

Parabolic intersections in Artin groups

OpenAI

  • Mathematics
  • Theoretical Computer Science

The September 23 manuscript claims that the intersection of any two conjugates of standard parabolic subgroups is parabolic in every finite-rank Artin group, including nonspherical types and infinite Coxeter labels.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

The Margulis–Platonov conjecture over number fields

OpenAI

  • Mathematics

For an absolutely almost simple simply connected group G over a number field k, every noncentral abstract normal subgroup of G(k) is the inverse image of an open normal subgroup of the product of G(kᵥ) over anisotropic nonarchimedean places.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Three Conjectures on Binary Channels for the Doubly Symmetric Binary Source

Georg Pichler

  • Mathematics
  • Theoretical Computer Science

For a doubly symmetric binary source with crossover p∈[0,1], the convex hull of rate triples from arbitrary binary channels equals that from binary symmetric channels. At p=0 and parameters a,d∈(0,1), no binary channel pair with rates no larger than the stated Z/S pair has greater I(U;V), and no pair with rates at least those of the stated Z/Z pair has smaller I(U;V). These are value bounds, not uniqueness claims.

Integer multiplication below n log n

OpenAI

  • Mathematics
  • Theoretical Computer Science

The manuscript reports deterministic multiplication in O(n (log n)^(1−κ)) time with κ = 2^−182 on a fixed finite-tape Turing machine.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Aperiodic maximisers of Lyapunov dimension: A counterexample to the unrestricted form of Eden's conjecture

Matthew Colbrook

  • Mathematics

For each c>4, Colbrook’s polynomial ODE on ℝ⁵ has uniformly negative divergence and a compact global attractor. Its finite-time local Lyapunov and asymptotic Kaplan–Yorke dimension maxima are 4+4/c on an irrational-flow invariant two-torus, while equilibrium and periodic-point maxima are 3.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Small undecidable groups and fixed-size Adian–Rabin families

Marc Kegel, Shana Yunsheng Li, Qiuyu Ren

  • Mathematics
  • Theoretical Computer Science

Kegel, Li and Ren give an effective family of finite presentations with two generators and ten relators for which deciding triviality is impossible. The count is an achieved presentation size, not a minimum.

AI involvementAI-led
ImportanceNotable
Human understandingFull

Small undecidable groups and fixed-size Adian–Rabin families

Marc Kegel, Shana Yunsheng Li, Qiuyu Ren

  • Mathematics
  • Theoretical Computer Science

Kegel, Li and Ren give an effective family of finite presentations with four generators and eleven relators for which deciding triviality is impossible. These are fixed presentation slots, not a minimality result.

AI involvementAI-led
ImportanceNotable
Human understandingFull

AI-assisted certificate reports optimality of logarithmic-spiral shoreline search

  • Mathematics
  • Theoretical Computer Science

Alexander Temerev reports that every planar search path for an unknown straight shoreline has competitive ratio at least 13.8111351794611…, matching the logarithmic spiral. The proof permits paths with decreasing radius or reversing polar angle.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

On Kahn's flow conjecture

Peter Keevash

  • Mathematics

For every increasing antipodal function h on the Boolean cube and every nonnegative allocation of each nonempty Fourier coefficient’s squared mass among its coordinates, the induced linear function flows upward to h₊². This is Kahn’s flow conjecture.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

A cyclic non-Ramsey heptagon

Dömötör Pálvölgyi

  • Mathematics

For every transcendental radius r>2, a seven-point set on a circle of radius r, including r=π, has a finite coloring of Euclidean space with no monochromatic congruent copy in any dimension.

AI involvementAI-autonomous
ImportanceMajor
Technical verificationReported
Human understandingFew

AI-assisted construction refutes a Naslund bound over F₃

  • Mathematics
  • Theoretical Computer Science

A ten-word code in F₃⁴ and an eight-degree lift give square-difference-free polynomial sets of size 810ᵉ below degree 8e and 27·810ᵉ below degree 8e+4. They refute Naslund’s Conjecture 13 at q=3, k=2 for every n divisible by four with n≥8, and imply a growth-exponent lower bound of 16/21.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

Proof of the Pach-Tardos conjecture

Lior Gishboliner, Xiangyu Li

  • Mathematics
  • Theoretical Computer Science

For each fixed acyclic 0–1 matrix pattern P, Gishboliner and Li prove Ex(n,P)≤n^(1+O_P(1/log log n)), establishing the weak n^(1+o(1)) Pach–Tardos statement.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFull

Astra-origin construction gives discontinuous traces

  • Mathematics

Sauers constructs a separable unital C*-algebra with compact nonempty trace space whose uniform tracial completion has a trace discontinuous in the uniform 2-norm, and a factorial tracially complete pair with the same failure.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Codex proves two arithmetic-function conjectures in Lean

  • Mathematics
  • Theoretical Computer Science

Lean proofs establish Dimitrov’s Conjectures 1 and 3, the stated rational lower bounds involving Euler’s totient φ(n), Dedekind’s ψ(n) and σ(n), for every integer k≥1 and n≥2. Equality in each bound holds exactly when n is prime.

AI involvementAI-led
ImportanceModest
Human understandingFew

An Infinite Simple Group with Property (T) That Is Sofic and Hyperlinear

Sauers

  • Mathematics
  • Theoretical Computer Science

A group built as EL₃(LC(X,F₂)⋊ℤ), with X a binary Toeplitz subshift, is infinite and simple, has Kazhdan property (T), and is locally embeddable into finite groups. It is therefore sofic and hyperlinear. The Lean project compares two existential statements using this same witness.

AI involvementAI-led
ImportanceMajor
Human understandingFew

Confinement schemas and the reach of block certificates for powers of rational numbers modulo one

Ralf Stephan

  • Mathematics

For every nonzero real ξ, the fractional parts of ξ(3/2)ⁿ cannot all lie in the closed interval [2/7,5/7]. Equivalently, ||ξ(3/2)ⁿ||<2/7 infinitely often. The certificate has depth 15 and 166 blocks across 32 rank strata.

A Tight Upper Bound for Fully Dynamic Matching on the Line

Yash Kanoria

  • Mathematics
  • Theoretical Computer Science

For m≥2 labeled supplies on [0,1], independent uniform demands and independent uniform replacement after each online match, the project proves a policy whose one-period root-mean-square cost, from every fixed initial inventory, has limsup at most (72000/log 2)·log(m+1)/m. After a scheduled m-period replacement phase, expected average cost including that phase obeys the same bound for every horizon N≥2m².

AI involvementAI-led
ImportanceNotable
Human understandingFull

Erdős-Sós for digraphs

Dhruv Mubayi, Jacques Verstraete

  • Mathematics
  • Theoretical Computer Science

Every finite loopless Eulerian digraph with no repeated ordered arc, including digraphs with opposite arcs, contains every orientation of every t-edge tree when it has n vertices and more than (t−1)n arcs. Complete bidirected graph unions show the bound is sharp.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFull

The uniform Turán density of the tetrahedron

Matija Bucić

  • Mathematics
  • Theoretical Computer Science

In Bucić’s alternative proof that the uniform Turán density of the three-uniform tetrahedron K₄³ is 1/2, ChatGPT 6 Pro found the finite sum-of-positive-quadratic-forms identity and proposed the tensor projection inequality used in the proof.

A counterexample to the periodicity conjecture for finite-dimensional algebras

Haruhisa Enomoto

  • Mathematics

Enomoto constructs a 36-dimensional algebra Λ(q) over an algebraically closed field with an element q of infinite multiplicative order. Every simple right module has period four and the algebra is twisted periodic, yet a two-dimensional module has pairwise nonisomorphic even syzygies, so the algebra is not periodic.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFull

The bounds ℏₚ,ₓ ≲ βₚ,ₓ² and βₚ,ₓ ≲ ℏₚ,ₓ² are sharp

Emiel Lorist, Jan van Neerven

  • Mathematics

Lorist and van Neerven construct two explicit 2^n-dimensional Banach-space families at p=2. One has Hilbert-transform constant of order n and UMD constant of order √n; the other reverses them, making both classical quadratic comparison exponents sharp over real and complex spaces.

AI involvementAI-centered
ImportanceMajor
Human understandingFull

The ceiling of the single-letter entropy method for the union-closed sets conjecture, and a protocol that reaches it

Andrew Moffat

  • Mathematics
  • Theoretical Computer Science

Moffat’s Theorem 3.1 claims to bound any single-letter entropy certificate whose classes contain product laws by c<=1-h(1/sqrt(2))/sqrt(2)=0.383099... . Theorem 3.4 gives a refined c<=0.382885260... ceiling when the iid protocol has a singleton class and other classes admit component hiding. These are limits of specified certificate methods, not bounds on the true union-closed constant. The paper's c>=0.38284 construction remains conditional on numerical hypotheses and is separate.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Finite time blowup for Navier–Stokes

OpenAI

  • Mathematics

For every positive viscosity, the manuscript reports a three-dimensional incompressible Navier–Stokes flow starting from rest with smooth compactly supported forcing, bounded kinetic energy, and unbounded velocity in finite time. It claims Clay alternatives C on R³ and D on the periodic torus.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Finite time blowup for the Euler equation

OpenAI

  • Mathematics

The manuscript reports smooth compactly supported divergence-free initial data on R³ whose unforced three-dimensional incompressible Euler solution has finite maximal lifespan, with unbounded velocity gradient and divergent time integral of maximum vorticity.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFew

Kalai's Conjecture for Tight Trees

Dhruv Mubayi, Jacques Verstraete

  • Mathematics

For n≥r≥2, t≥1 and any n-vertex r-uniform hypergraph H containing no tight r-tree T with t edges, |E(H)|≤(t−1)|∂H|/r≤(t−1)binom(n,r−1)/r, where ∂H is the (r−1)-shadow. The bound is attained for infinitely many n.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Lean-checked candidate against the φ-mixing CLT

  • Mathematics

GPT-6 Astra constructed a strictly stationary, centered φ-mixing sequence with finite second moments and divergent partial-sum variance. Along a subsequence, its variance-normalized sums converge to zero in probability rather than to a standard normal law.

The optimal universal exponential base for Euclidean kissing numbers

Sébastien Palcoux

  • Mathematics

For every positive dimension n, the Euclidean kissing number τₙ is an attained maximum and satisfies τₙ≤(√6)ⁿ. A planar hexagon gives τ₂=6, so √6 is the least base in a bound τₙ≤αⁿ valid in every dimension without a prefactor.

AI involvementAI-led
ImportanceNotable
Human understandingFew

A counterexample to Petrykowski's conjecture

Artem Chernikov

  • Mathematics
  • Theoretical Computer Science

Chernikov constructs a complete simple theory with a definable group and a global type whose left-translation orbit has size at most 2^ℵ₀, although the group has no left-invariant Keisler measure. This refutes the bounded-orbit implication to definable amenability outside NIP.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

GPT-6 Astra disproves Erdős’s sum-distinct lower-bound conjecture

  • Mathematics
  • Theoretical Computer Science

For every ε>0, arbitrarily large n admit an n-element sum-distinct set of positive integers with maximum element N≤ε·2^n. The formal statement excludes N=0 and gives no effective n as a function of ε.

AI involvementAI-autonomous
ImportanceNotable
Human understandingPartial

GPT-6 Astra proves a superlogarithmic bound for Erdős Problem 126

  • Mathematics
  • Theoretical Computer Science

Let f(n) be the minimum number of distinct prime divisors among pairwise sums of any n-element subset of positive integers. The compared theorem proves f(n)/log n→∞. Separate Lean resolutions give stronger polynomial lower bounds, including a square-root bound.

AI involvementAI-autonomous
ImportanceNotable
Human understandingPartial

GPT-6 Astra resolves Erdős Problem 74 on local bipartiteness

  • Mathematics
  • Theoretical Computer Science

There is a divergent integer-valued function f such that if every finite n-vertex subgraph of a graph becomes bipartite after at most f(n) edge deletions, then the graph has finite chromatic number. The formal proof gives three-colourability.

AI involvementAI-autonomous
ImportanceNotable
Human understandingPartial

AI-built Lean proof reports Conway’s omnific refinement property

  • Mathematics
  • Theoretical Computer Science

For cut-defined omnific integers a,b,c,d represented as numeric games, a Lean theorem says ab=cd implies omnific e,f,g,h with a=ef, b=gh, c=eg and d=fh, including zero cases. The theorem is stated at each Lean universe, rather than putting all surreal numbers into one type.

AI involvementAI-led
ImportanceNotable
Human understandingFew

GPT-6 Astra proves the rational-exponents conjecture

  • Mathematics

For every rational r with 1 ≤ r < 2, there is a finite bipartite graph H whose extremal number ex(n,H) grows as Θ(n^r). Here ex(n,H) is the largest number of edges in an n-vertex graph containing no copy of H; Θ means matching upper and lower bounds up to constant factors.

AI involvementAI-autonomous
ImportanceMajor
Technical verificationFormal (Lean)
Human understandingFull

A machine-checked bound of 0.6934 for Besicovitch's 1/2-problem

Yongxi Lin

  • Mathematics

For every real Hilbert space H, including nonseparable spaces, the one-dimensional Besicovitch rectifiability threshold σ₁(H) is at most 0.6934. In particular, a planar Borel set of finite H¹ measure is countably one-rectifiable when its lower one-density exceeds any fixed threshold greater than 0.6934 almost everywhere.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Interior Curvature Estimates for the Graphical Scalar Curvature Equation in All Dimensions

Guohuan Qiu, Jin Yan

  • Mathematics

Qiu and Yan prove interior curvature bounds in every dimension n≥3 for smooth admissible graphs on B₃⁄₂ solving σ₂(κ[u])=1 with bounded height and gradient. The model-suggested compression inequality in Lemma 3.4 controls a two-surface Jacobi inequality; the model also supplied the auxiliary η in equation 62.

AI involvementAI-centered
ImportanceMajor
Technical verificationReported
Human understandingFull

Optimal girth-dependent bounds for the Bethe approximation of the permanent

Dingding Dong, Vishesh Jain

  • Mathematics
  • Theoretical Computer Science

Dong and Jain report that, for a nonnegative n-by-n matrix whose bipartite support graph has girth at least even g>=4, Theorem 1.3 proves Bethe(A)<=per(A)<=2^(2n/g)Bethe(A); disjoint unions of g-cycles attain the upper bound when g divides 2n. The model-developed intermediate proof had a weaker exp(n Theta(log g/g)) factor; the exact model-origin step retained in the final sharp proof is not identified.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

A quantum oracle separation between QMA(2) and QMA

John Bostanci, Sabee Grewal, Jonas Haferkamp, Andrew Huang, Yeongwoo Hwang, Anand Natarajan, Chinmay Nirkhe

  • Mathematics
  • Theoretical Computer Science

Bostanci and coauthors construct a unitary quantum oracle U for which QMA(2)^U differs from QMA^U. Their oracle problem has a one-query QMA(2) protocol with linear-size unentangled proofs, while QMA needs exponential resources in query count or proof size.

AI involvementAI-centered
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

The entropy production is not always monotone for hard spheres or for Maxwell molecules

Luis Silvestre

  • Mathematics
  • Theoretical Physics

For the three-dimensional space-homogeneous Boltzmann equation with collision kernel B=|v−v*|^γ/(4π), 0≤γ≤1, Silvestre constructs smooth positive radial two-Maxwellian initial densities f_R=(1−p)M₁+pM_R. For 0<p<p_γ and sufficiently large R, the entropy production D initially increases: dD(f_R)/dt>0 at t=0, with a positive leading term proportional to R^γ log R. The range includes constant-angle Maxwell molecules and hard spheres.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFull

AI-led Lean proof gives a quadratic orbit-gap criterion

  • Mathematics

For a quadratic real root α>1 of X²−aX−b whose conjugate has modulus below one, the Route A exponent condition A(α)<1 gives an open interval missed by every orbit (ξαⁿ mod 1) with ξ in the paper’s class C(α), so none is uniformly distributed. The criterion is equivalent to (log₂α−1)(log₂(α/|b|)−1)>1 and yields an explicit gap for α=2+√5.

AI involvementAI-led
ImportanceModest
Human understandingFew

AI-assisted proof settles the odd-exponent Colombo case

  • Mathematics

For even N=2m≥2, pairwise distinct real numbers x₁,…,x₂ₘ and odd D=2r+1≥N−1, Ma proves that the matrix [(xⱼ−xᵢ)^D] is nonsingular. After sorting the nodes, the proof gives the stronger sign (−1)^(m choose 2) Pf[((xⱼ−xᵢ)^D)]>0.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

A counterexample to Scottish Book Problem 155

Yoshito Ishiki

  • Mathematics

Yoshito Ishiki constructs real Banach spaces X and Y and a bijection U from X to Y that preserves distances between every pair of points in each closed ball of radius 1/4, but is not a global isometry. This answers the unrestricted Scottish Book Problem 155 negatively.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Local limits along squares and prime values of digital functions

  • Mathematics
  • Theoretical Computer Science

For every base b≥2 and sufficiently large digit sum q congruent to a square modulo b−1 (automatic for b=2), the authors construct at least exp(c_b√q) squares n² of bounded size with gcd(n,b)=1 and base-b digit sum q. The explicit Lean statement gives q≥2304b⁴, n²≤b^(3q), and at least 2^(⌊√q/(36b)⌋) examples for b≥3, with a separate binary case.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

AI-assisted Lean proofs give a hard-sphere volume identity and bound

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

For k≥2 in any positive integer dimension, the magnitude of the kth hard-sphere Mayer coefficient, multiplied by k!, equals a sum of no-broken-circuit region volumes over spanning trees. In dimension three, each tree-region volume is at most (17/32)^ν(T)κ^(k−1), where κ is unit-ball volume and ν(T) is the largest packing of disjoint ordered forks a<b<c with tree edges ab and ac.

AI involvementAI-centered
ImportanceNotable
Human understandingFew

Improving Randomized Metric Distortion to 2.1441

Nisarg Shah

  • Mathematics
  • Theoretical Computer Science

For every finite metric voting profile, Shah’s v2 random-size stable-lottery rule has distortion at most 107200742653/50000000000 = 2.14401485306. This supersedes the v1 mixed-rule bound of 2.3282.

Non-MF groups and non-finite full group C*-algebras

Caleb Eckhardt

  • Mathematics

From a property-(T) group Γ with an injective non-surjective endomorphism, Eckhardt forms its ascending HNN extension G and generalized wreath product W=(⊕_(G/Γ) Z/2Z)⋊G. His Theorem 4.3 gives a specified nonidentity element killed by every homomorphism of W to a matrix-ultraproduct quotient, showing W is not MF.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFull

On supporting affine functionals for Entanglement of Formation

A. S. Holevo, M. E. Shirokov

  • Mathematics

For the two-qubit state ρ=(|Φ+⟩⟨Φ+|+|01⟩⟨01|)/2, where |Φ+⟩=(|00⟩+|11⟩)/√2, Holevo and Shirokov show that entanglement of formation has no global supporting affine functional. Along a perturbation toward |10⟩, it decreases on a square-root scale.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

AI-generated construction raises a square-difference-free lower bound

  • Mathematics

An eleven-block construction yields α∞≤liminf log D(N)/log N for the largest square-difference-free subset D(N) of {1,…,N}. Every exponent ρ<α∞ holds eventually, and the formal proof certifies α∞>0.7537.

AI involvementAI-led
ImportanceModest
Human understandingFew

AI-found graph refutes Graffiti Conjecture 806 on a distinct-degree reading

  • Mathematics
  • Theoretical Computer Science

For PR[S] on square-free integers S⊂[2,n], with edges where gcd exceeds one, the n=51 graph has an exact integer Rayleigh witness showing its largest adjacency eigenvalue exceeds its number of distinct degree values.

AI involvementAI-led
ImportanceModest
Human understandingFew

A Spectral Local-to-Global Principle for Spin Systems on Graphs with Girth At Least Five

Xiaoyu Chen, Kuikui Liu

  • Mathematics
  • Theoretical Computer Science

For each δ∈(0,1), sufficiently large maximum degree Δ and finite simple n-vertex graphs of girth at least 5, Theorem 1 gives irreducible proper-q-coloring Glauber dynamics with spectral gap Ωδ(1/n) and mixing time Oδ(n² log q+n log(1/ε)) when q≥(1+δ)Δ. Theorem 3 extends the method to antiferromagnetic Potts models under its stated β and q conditions; Theorem 5 gives a local spectral-contraction criterion for general multi-spin systems. The result is rapid mixing, not the conjectured exact q≥Δ+2 threshold or uniformly optimal O(n log n) bound.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Three omitted values and non-Blaschke point divisors in half-planes

Quanyu Tang, Bokai Cui, Wei He, Tao Hu, Yanyang Li, Ke Wang, Zijun Yu

  • Mathematics

Tang et al.’s Theorem 1.2 constructs a real meromorphic F on C with all 0-, 1- and ∞-points on the real axis, each infinite, and for every other spherical value a, the a-point divisor in each half-plane fails the Blaschke condition. Consequently F is not of bounded type in either half-plane despite omitting three prescribed values there; Möbius postcomposition and affine precomposition extend the example to any triple and Euclidean half-plane.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationPartial
Human understandingFew

Sparse domination implies convex body domination

Aapo Laukkarinen, Emiel Lorist

  • Mathematics

The AI-associated dyadic randomization step underlies Theorem A: for 1≤r,s<∞ with 1/r+1/s>1, a bilinear form on compactly supported bounded scalar functions with an (r,s)-sparse bound has a convex-body sparse bound for its coordinate-wise C^n extension, with stated n^(1+1/r+1/s) and η^−1 dependence. Theorem 2.6 proves a dyadic version; the non-dyadic extension and final theorem were completed by the human authors.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Fröberg's Conjecture for Quintics and Septics in Four Variables

Qihang Wang, Dongming Zhang

  • Mathematics

Over any characteristic-zero field, Wang and Zhang prove Fröberg’s predicted generic Hilbert series for equal-degree forms in four variables when the degree is 5 or 7 and the generator count is any r≥1. Exact rank certificates cover the previously open middle ranges r=6–11 and r=6–21, respectively; propagation covers the remaining counts.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFew

A Discontinuous Solution of the Critical n-Laplace System with Antisymmetric Potential

Dominik Schlagenhauf

  • Mathematics

For n>2, Schlagenhauf constructs a bounded W¹,ⁿ map from the n-ball to Rⁿ⁺² that is discontinuous at the origin yet solves the critical n-Laplace system with an antisymmetric Lⁿ potential in distributions. The potential can lie in every L⁽ⁿ,ᑫ⁾ for q>2.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Generically stable Keisler measures

Gabriel Conant, Kyle Gannon, James E. Hanson

  • Mathematics
  • Theoretical Computer Science

For Borel-definable global Keisler measures over a model of a discrete first-order theory, the authors prove the reverse implications: self-averaging gives a definable measure whose canonical random extension is generically stable, which in turn gives a frequency interpretation measure. Together with earlier forward implications, these yield equivalence.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFull

Entropy of Bernoulli Measures Conditioned on Affine Subspaces and a Problem of Ancheta--Massey

Yihong Wu

  • Mathematics
  • Theoretical Computer Science

For iid Bernoulli(p) bits with 0<p≤1/2 and a full-row-rank k×n binary linear encoder H, its optimal bitwise-MAP distortion D(H) obeys k/n≥h(p)(1−D(H)/p). A nonasymptotic entropy bound for every affine subspace drives the converse.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Autonomous Mathematical Discovery in an Open-World Multi-Agent Environment

Stephen Chung, Wenyu Du, William J. Wesley

  • Mathematics

Station v2 Section 6.13 S1: a conference graph with parameters (q,(q-1)/2,(q-5)/4,(q-1)/4) gives R(B_q,B_(q+1))=4q+3. S2: every prime power Q>3 with Q=3 mod 8 gives R(B_((Q-1)/2),B_((Q+1)/2))=2Q+1. Exclude the third Yamada–Pott family requiring human expert synthesis.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean)
Human understandingFull

A Sparse-Fock Proof of the Upper Edge for Fully Independent SparseStack

Diar Heidary

  • Mathematics
  • Theoretical Computer Science

For a fixed n×d real matrix U with UᵀU=I, d≥1, 0<ε≤1 and 0<δ≤1, the fully independent signed-hash SparseStack construction gives failure probability at most δ with column sparsity below 1356 max(1,⌈log₂(d/δ)⌉)/ε and row count below 306456(d+max(1,⌈log₂(d/δ)⌉))/ε².

AI involvementAI-centered
ImportanceNotable
Human understandingFull

The (3, 4, ∞) modular family of 2-tori, completed at its three special points, is a complex structure on S⁶.

  • Mathematics

The Alpöge–Claude manuscript claims a compact complex threefold, obtained by completing a modular family of complex two-tori, whose underlying smooth manifold is diffeomorphic to S⁶. Philip Engel supplies a separate geometric exposition and self-contained proof of the construction.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Blowup for the Euler Equations with Smooth Forcing

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

Alpöge and Buckmaster construct smooth compactly supported initial velocity and forcing smooth through blowup time for 3D incompressible Euler on R³. A finite-energy classical solution develops unbounded vorticity in finite time, with divergent time-integrated vorticity sup norm and uniqueness in the stated preterminal class.

An improved algebraic construction for Ramsey numbers

Ferdinand Ihringer, Sam Mattheus

  • Mathematics

ChatGPT 5.6 suggested the two-way trace graph TG_{d,h} of Remark 2.2 on projective points over F_(2^h). Ihringer and Mattheus bound its clique size by binom(d−2,ceil((d−2)/2)) and independence number by (d^h+1)². Their stronger one-way ordered graph and Theorem 1.3 are human-developed refinements.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

A counterexample to the bounded mass property

Mingchen Xia, Kewei Zhang

  • Mathematics

For the Hopf threefold X=(C³\{0})/⟨z↦e^-1 z⟩, Xia and Zhang’s Theorem 1.1 constructs a smooth Hermitian form ω and smooth potentials φ_j with ω+dd^cφ_j>0 yet ∫_X(ω+dd^cφ_j)^3→∞. Thus this compact complex manifold lacks the bounded mass property. The paper’s proof uses an elliptic fibration, an exact mass identity and fiberwise heat-kernel potentials.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFull

Even integral parts of powers of quadratic Pisot numbers

Ralf Stephan

  • Mathematics

Among real quadratic Pisot numbers α, the golden ratio is the only one in Dubickas’s set S: every other α admits a nonzero ξ for which all floors ⌊ξαⁿ⌋ are even for n≥1. For α=1+√2, ξ=(2−√2)/4 gives the exact Pell identity ⌊ξ(1+√2)⁽ⁿ⁺¹⁾⌋=Pₙ−[n odd].

AI involvementAI-led
ImportanceNotable
Human understandingFew

AlphaProof Nexus finds a formal cuboid irreducibility proof

  • Mathematics
  • Theoretical Computer Science

For positive coprime unequal integers a and b, an AlphaProof Nexus Lean proof shows that Sharipov’s degree-eight cuboid polynomial Pₐ,ᵦ(t) is irreducible over the integers. Its route uses quartic factor analysis and an infinite descent for x⁴+34x²y²+y⁴=z².

AI involvementAI-autonomous
ImportanceModest
Human understandingFull

An AI-discovered smooth random fast dynamo on T³

Keefer Rowan

  • Mathematics
  • Theoretical Physics

For the explicit iid block-refreshed smooth divergence-free velocity field U on T³ and the specified divergence-free initial magnetic mode, Theorem 1.1 gives, for each κ in [0, log(π/e)/(4π²)], an almost surely finite κ-dependent L^κ with ||B_t^κ||₂≥(L^κ)^−1 e^(t/2) for all t≥0 and a uniform positive moment bound on L^κ. Null sets can depend on κ; no single deterministic field working simultaneously for every κ is proved.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Sub-optimality of Marton's Inner Bound for the Two-Receiver Broadcast Channel

Mian Huang, Yanxiao Liu, Yi Liu

  • Mathematics
  • Theoretical Computer Science

Huang, Liu and Liu give a finite two-receiver discrete memoryless broadcast channel T for which Marton’s two-use sum-rate quantity M(T⊗2) exceeds twice its one-use value M(T), with a certified gap above 1.8821×10⁻⁶ nats. Its one-letter Marton region is therefore strictly smaller than capacity.

Erdős Problem 501 after adding ω₂ random reals

Elliot Glazer, Sol

  • Mathematics
  • Theoretical Computer Science

A ZFC model has the property that every family (Aₓ) of subsets of ℝ with outer measure below one admits an infinite X such that x∉Aᵧ for distinct x,y∈X. Combined with Hechler’s earlier CH counterexample, this gives relative ZFC independence for the first question of Erdős 501.

Claude reports the general four-color Rado formula

  • Mathematics
  • Theoretical Computer Science

For four colors, with x=y allowed, the reported proof gives R(c)=40c+41 for every integer c≥2 in x+y+c=z. It reduces the universal statement to R(2)=121 and two unsatisfiable spoke templates.

AI involvementAI-led
ImportanceNotable
Human understandingFew

On a conjecture on the Kasami APN function: reductions, structure theorems, a proof for k mod n ∈ {1, 2, n−2, n−1}, and exhaustive verification for n ≤ 13

Gábor P. Nagy, Attila Vajda

  • Mathematics

For the Kasami function F(x)=x^(4^k−2^k+1) over F_(2^n), with gcd(k,n)=1, Nagy and Vajda prove the conjectured ordered triple count 2^(2n−3) for all distinct nonzero coefficients when k mod n is 1, 2, n−2 or n−1. They also check admissible n≤13 by exhaustive computation.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4 with Mathlib)
Human understandingFew

A 15/31 Counterexample Family to the Albertson-Berman Conjecture

Heejae Jung

  • Mathematics

Jung constructs a 31-vertex simple plane triangulation with maximum induced forest 15, then simple maximal planar graphs Mₖ with 31k vertices, minimum degree five and maximum induced forest 15k for every k≥2. The ratio 15/31 refutes the Albertson–Berman half-forest conjecture.

Improving the matrix multiplication exponent with modern optimization and AlphaEvolve

Emilien Dupont, Marvin Eisenberger, Borislav Kozlovskii, Abbas Mehrabian, Francisco J. R. Ruiz, Abigail See, Renfei Zhou, Josh Alman, Virginia Vassilevska Williams, Matej Balog

  • Mathematics
  • Theoretical Computer Science

Dupont and collaborators report ω<2.371177 by improving the combination-loss optimization and refining it with AlphaEvolve.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

Counterexamples to Makeev's regular-simplex conjecture in dimensions four and five

ChatGPT

  • Mathematics

In each of dimensions four and five, for every sufficiently small nonzero ε, an explicit smooth odd function on S^(d−1) cannot agree with a linear functional on any rotated normalized A_d root system. Through Kuperberg’s equivalence, these give counterexamples to Makeev’s universal-cover conjecture in those dimensions.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Weak-Type Bounds for Convolution on the Boolean Hypercube

Junwei Lu, Shengtao Guo, Ethan X. Fang

  • Mathematics

On the Boolean cube {−1,1}^n with uniform measure, for every fixed 0<a<1, n≥1 and u>1, convolution by the biased-coin product measure μ_a satisfies sup_{f≥0, ||f||₁=1} u·λ({T_{μ_a}f≥u})≤C_a/√(log u), with an a-dependent constant. The bound is dimension free.

AI involvementAI-centered
ImportanceMajor
Technical verificationReported
Human understandingFew

AI-Assisted Discovery and Construction of a Counterexample to the Convergence of Three-Block ADMM with the Identity Matrix as its Third Constraint Block

Kenan Xu, Xiangfeng Wang

  • Mathematics
  • Theoretical Computer Science

Xu and Wang exhibit a strongly convex rational quadratic program with A=B=I₂ and identity third block whose direct three-block ADMM has a bounded non-KKT orbit of minimal period 66. A separate Kimi search gives a locally attracting period-23 orbit in dimension three that can be converted to all-identity blocks.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFull

Unbounded variation solutions for uniformly elliptic equations in nondivergence form in dimension three

Nam Q. Le, Qi Sun, Hung V. Tran

  • Mathematics

Le, Sun and Tran construct smooth three-dimensional uniformly elliptic nondivergence equations with I≤Aₘ≤2⁸¹I, common boundary data and bounded solutions whose gradient L¹ norms diverge. They also obtain a uniformly convergent approximation limit that is not locally of bounded variation.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

A counterexample to the Foregger-Sinkhorn tie-point conjecture

Yair Lavi

  • Mathematics
  • Theoretical Computer Science

An explicit 8×8 nearly decomposable support has a unique global permanent minimizer on its full doubly stochastic face. At a specified zero, the permanental cofactor exceeds the permanent although that zero is not a tie point, refuting the Foregger–Sinkhorn implication.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

A solution to Banach's isometric conjecture

Xinbao Lu, Kaiwen Yang

  • Mathematics

For odd n≥3, Lu and Yang obtain a global Lipschitz family of exact hyperplane-section isometries after a positive-degree bundle pullback. Combined with their degree argument, this proves that a real Banach space whose n-dimensional subspaces are all linearly isometric is Hilbert when 2≤n<dim X.

AI involvementAI-centered
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

SOP₂ = SOP₃

Artem Chernikov

  • Mathematics
  • Theoretical Computer Science

For every complete first-order theory, SOP₂ implies SOP₃. The reverse implication was already known, so the two dividing lines coincide.

AI involvementAI-led
ImportanceMajor
Technical verificationReported
Human understandingFull

A dense-case theorem for Seymour's second neighborhood conjecture

Jake Brukhman

  • Mathematics

Every oriented graph on n=2δ+2 vertices with minimum outdegree δ has a vertex whose exact second outneighborhood is at least as large as its first. With Fisher’s tournament theorem, this covers n≤2δ+2, without proving Seymour’s conjecture for all oriented graphs.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFew

Grothendieck's theorem for Bessel sequences

Lukas Liehr, Mitchell A. Taylor, Peiyang Yu

  • Mathematics
  • Theoretical Computer Science

Liehr, Taylor and Yu prove that every countable Bessel sequence of bound 1 in a complex Hilbert space has its full Gram matrix, including diagonal entries, represented by correlations of complex functions on [0,1] bounded in modulus by 1.

AI involvementAI-centered
ImportanceMajor
Human understandingFull

No compromise in the liquid drop model

Otis Chodosh, Matilde Gianocca

  • Mathematics

For the three-dimensional liquid-drop energy given by De Giorgi perimeter plus half the Coulomb self-interaction, balls uniquely minimize at every fixed volume 0<V≤5(2−2^(2/3))/(2^(2/3)−1)≈3.51, up to translation and null sets. Above that threshold, no minimizer exists.

AI involvementAI-autonomous
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Lions' Maximal Regularity Problem for Divergence-Form Differential Operators: Failure at the 1/2-Hölder Endpoint

Lukas Niebel

  • Mathematics

Niebel constructs a real scalar diffusion coefficient on a bounded interval that is bounded, uniformly elliptic, one-half Hölder in time with spatial L∞ values, and arbitrarily close to the heat coefficient, yet its zero-data variational solution with continuous L² forcing has a time derivative outside space-time L². Extensions give real symmetric isotropic counterexamples in all dimensions and bounded domains.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Solutions to Five Challenge Problems in Enumerative and Algorithmic Combinatorics, with an Account of the Human-Machine Methodology Employed

Jaideep Sai Padhi

  • Mathematics

For the two-layer solid standard Young-tableau shape [[n,n],[n,1]], Padhi gives an explicit two-term hypergeometric formula for the count g(n) and proves the conjectured second-order degree-12 recurrence for every n≥1.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Solutions to Five Challenge Problems in Enumerative and Algorithmic Combinatorics, with an Account of the Human-Machine Methodology Employed

Jaideep Sai Padhi

  • Mathematics
  • Theoretical Computer Science

For every fixed r,s≥1, both restricted-permutation count sequences a_(r,s)(n), forbidding π_(i+r)−π_i=s, and b_(r,s)(n), forbidding the corresponding absolute difference, are claimed to be holonomic. The argument uses an interleaving bijection with finitely many seams and D-finite closure.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Solutions to Five Challenge Problems in Enumerative and Algorithmic Combinatorics, with an Account of the Human-Machine Methodology Employed

Jaideep Sai Padhi

  • Mathematics

For three-row tableaux with runs of length at least two, G(n)~C₁8ⁿ/n⁴; for odd runs, H(n)~C₂(7+5√2)ⁿ/n⁴. The constants are positive finite boundary sums, with numerical estimates about 0.52128605909 and 0.6389278129.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

Solutions to Five Challenge Problems in Enumerative and Algorithmic Combinatorics, with an Account of the Human-Machine Methodology Employed

Jaideep Sai Padhi

  • Mathematics
  • Theoretical Computer Science

For the Geode G defined by S−1=(t₂+⋯+t₆)G, a coefficient-collapse identity reduces its diagonal coefficient to a one-dimensional integral and coefficient extraction. It enables an exact 8,367-digit computation of G[1000^(5)] using modular arithmetic and bounded Chinese remainder reconstruction.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFew

A lower bound for stepsize-based acceleration of gradient descent

Jianhao Ma, Yuxin Chen

  • Mathematics
  • Theoretical Computer Science

For every p between sqrt(2+sqrt(3)) and 2, every horizon T and every predetermined nonnegative gradient-descent stepsize schedule, Ma and Chen construct a smooth convex problem in dimension at most T+1 with last-iterate error at least c_p L R²/(T+1)^p.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4 + human)
Human understandingFull

A Tight Lower Bound for Smooth Nonconvex Stochastic Optimization with Bounded Gradient Noise

Jikai Jin

  • Mathematics
  • Theoretical Computer Science

In the paper’s small-ε regime, a randomized adaptive method using one independent fresh sample per query needs Ω(ΔL/ε² + ΔLσ²/ε⁴) queries to reach expected gradient norm at most ε for smooth nonconvex objectives with gap Δ, L-Lipschitz gradient and uniformly bounded, unbiased fixed-law gradient noise.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFew

The sharp SAT/UNSAT phase transition in random ellipsoid fitting

Theodor Misiakiewicz, Garrett G. Wen

  • Mathematics
  • Theoretical Computer Science

For n independent standard Gaussian points in R^d, a positive semidefinite S with x_iᵀSx_i=d for every i exists with high probability when limsup n/d²<1/4 and fails to exist with high probability when liminf n/d²>1/4.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Codex reformulates graviton cut tests as root-cone flows

  • Mathematics
  • Theoretical Physics

Kehoe’s working preprint recasts global cut conditions in the general-region half-collinear single-minus graviton recursion as positive directed-tree flows and open-root-cone membership. Retarded vertices become weighted sums over feasible spanning trees; directed Matrix-Tree determinants apply in chambers with a full arborescence family. It classifies four- and five-block chambers and gives one non-decay five-graviton identity.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFew

Halpern Iteration Achieves Õ(ε⁻¹ᐟᵖ) pth-Order Oracle Complexity for Monotone Variational Inequalities

Lesi Chen, Xinliang Zhang, Hengyu Wang, Chengchang Liu, Yongchao Chen, Jingzhao Zhang

  • Mathematics
  • Theoretical Computer Science

Under the paper’s pth-order smoothness and oracle assumptions, the final Halpern-based method for smooth monotone variational inequalities attains residual rate Õ(T^(−p)), or Õ(ε^(−1/p)) oracle complexity.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Two-copy distillability of one-copy-undistillable negative-partial-transpose states in every dimension

Gelo Noel M. Tabia, Kai-Siang Chen, Min-Hsiu Hsieh

  • Mathematics
  • Theoretical Physics

At canonical point C of the DiVincenzo family, an explicit Schmidt-rank-two tight-frame witness proves one-copy undistillability but two-copy distillability for every local dimension d≥3.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Overlap distribution of the critical Sherrington-Kirkpatrick model

Hang Du, Brice Huang

  • Mathematics
  • Theoretical Physics

At inverse temperature β=1, the quenched law of N^(1/3) times the two-replica overlap converges to an explicit random Airy₁-based measure for both Ising and spherical Sherrington–Kirkpatrick models. Its second moment gives the limit of N^(2/3) E⟨overlap²⟩, answering Talagrand’s Conjecture 11.7.5.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

A Tight Bound for Facial Distance Patterns in Planar Graphs

Viktor Fredslund-Hansen, Shay Mozes, Oren Weimann

  • Mathematics
  • Theoretical Computer Science

Fredslund-Hansen, Mozes and Weimann report that, for a connected undirected unweighted planar graph with a designated face whose facial walk has k edges, the number of distinct facial distance-difference patterns is O(k^2), matching the known Omega(k^2) lower bound. The model supplied the forbidden-configuration and fixed-cardinality insight and a short prefix-trie proof; the paper's separate diameter algorithm is human-developed downstream work.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationReported
Human understandingFull

Approximating spin systems on planar graphs

Heng Guo, Xinyuan Zhang

  • Mathematics
  • Theoretical Computer Science

For planar graphs at sufficiently small fixed hard-core activity, including λ≤1/1443, Guo and Zhang give a randomized polynomial-time approximation scheme for the partition function without a maximum-degree bound.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Symmetry-guided constructions of absolutely maximally entangled states in five open cases

Samuel Bevins, Yunus Bidav

  • Mathematics
  • Theoretical Physics

Three explicit Hermitian self-dual MDS codes with parameters [12,6,7] over F25 and [18,9,10] over F121 and F169 establish AME(12,5), AME(18,11), and AME(18,13); one-party projection establishes AME(17,11) and AME(17,13). The two length-18 matrices come from nine-entry group-circulant kernels; the theorem asserts existence, not classification.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Fair and Efficient Balanced Allocations for Additive Valuations

Benjamin Cookson, Nisarg Shah, Paritosh Verma

  • Mathematics
  • Theoretical Computer Science

For any number of indivisible goods and agents with nonnegative additive valuations, an allocation exists with bundle sizes differing by at most one, envy-free up to one good, and no balanced fractional reallocation that Pareto improves it.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Linear dependence of time-frequency shifts of a Schwartz function

Markus Faulhuber, Philipp Petersen, Jordy Timo van Velthoven, Felix Voigtlaender

  • Mathematics

A nonzero Schwartz function has twelve distinct time-frequency shifts, eleven in one irrational translate of a half-integer lattice plus the origin, that are linearly dependent. The authors also prove a finite counterexample without a fixed twelve-shift count.

A Counterexample to Fourier Alignment in Single-Neuron Modular Addition

Gautam Neelakantan Memana

  • Mathematics
  • Theoretical Computer Science

For the single-neuron modular-addition setting, Appendix A extends Memana’s human counterexample to selection-independent Clarke trajectories and dead-neuron examples under the specified ReLU, dead-zone smoothing and fixed-step full-batch gradient-descent dynamics.

AI involvementAI-centered
ImportanceModest
Technical verificationReported
Human understandingFull

Maximally entangled states are not complete for pseudo-telepathy

Olivier Lalonde

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

Lalonde gives a bipartite inner-product game with four and three inputs and six outputs. It has a perfect strategy using a nonmaximally entangled state, but no perfect finite-dimensional strategy using a maximally entangled state.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

A Tight Bound on Online Vertex Cover under Edge Arrivals

Zhihao Gavin Tang, Yuhao Zhang

  • Mathematics
  • Theoretical Computer Science

For every deterministic fractional online vertex-cover algorithm and every ε>0, a finite bipartite edge-arrival instance forces a competitive ratio of at least 2−ε. The paper extends the lower bound to randomized integral or fractional algorithms against an oblivious adversary.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Scenario-Cover Geometry for a Four-Terminal Planar Unsplittable-Flow Gadget

Matthew Protti

  • Mathematics
  • Theoretical Computer Science

For the released fixed four-terminal planar acyclic gadget with coordinatewise positive cost-difference scenarios, the normalized additive upper arc-deviation supremum is (299−41√41)/32 with one scenario, 17/8 with two, 3 with three, and 4 with at least four. The two-scenario value is not attained under scenario-wise cost non-increase.

AI involvementAI-centered
ImportanceModest
Technical verificationPartial
Human understandingPartial

A numerical criterion for complex Hessian type equations on projective manifolds

Gao Chen, Sijie Nie, Yulun Xu

  • Mathematics

Chen, Nie and Xu’s cone-inclusion Lemmas 3.1 and 3.2 show that adjoining a linear factor preserves the required right-Noetherian structure and inclusion, while f−εf′ has a strengthened right-Noetherian property and inclusion. These lemmas feed a projective uniform numerical criterion for complex Hessian-type equations.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Impossibility of Perfectly Complete Many-Round Key Agreement in the QROM

Longcheng Li, Qian Li, Xingjian Li, Qipeng Liu

  • Mathematics
  • Theoretical Computer Science

In the Boolean-output quantum random oracle model, every finite-round perfectly complete key-agreement protocol with quantum local operations, independent initial private states and classical communication admits a passive eavesdropper that recovers the key with certainty using O((q_A+q_B)^5) classical oracle queries. The eavesdropper may use unbounded computation.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Asymptotically attaining the Moore bound

Wouter Cames van Batenburg, Samuel Korsky

  • Mathematics
  • Theoretical Computer Science

For every fixed positive diameter k, the maximum number nₖ(d) of vertices in a graph of maximum degree at most d and diameter at most k satisfies nₖ(d)/dᵏ→1 as d grows. Halved-flag graphs supply the construction.

AI involvementAI-centered
ImportanceMajor
Human understandingFull

A finite forbidden family with superlinear surplus and non-join extremal graphs

Chuandong Xu

  • Mathematics
  • Theoretical Computer Science

Xu constructs a fixed finite forbidden graph family with chromatic parameter p=2 and extremal surplus greater than c n^(3/2) for all sufficiently large n, yet with an extremal graph at each such order whose complement is connected.

AI involvementAI-led
ImportanceMajor
Human understandingFull

The Condition-Number Barrier in Sparse Least Squares

Honghao Lin, Vahab Mirrokni, David P. Woodruff

  • Mathematics
  • Theoretical Computer Science

Assuming the randomized rational-weighted exact-volume Small-Set Expansion Hypothesis, no randomized polynomial-time algorithm can guarantee both additive ε accuracy in rational sparse least squares and support O(k κₛ₊ₖ^(1−γ)) for any fixed γ∈(0,1]. The hard matrices have full column rank and at most two nonzeros per row.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Approximating two-terminal network reliability

Weiming Feng, Yucheng Fu, Heng Guo

  • Mathematics
  • Theoretical Computer Science

Feng, Fu and Guo give a fully polynomial randomized approximation scheme for s-to-t reachability probability in general directed graphs with independent edge-opening probabilities. An existing reduction extends it to undirected graphs.

AI involvementAI-centered
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics

A finite family of connected bipartite graphs, each containing a cycle, has joint extremal number O(n^(4/3−1/48)), while every member has extremal number Ω(n^(4/3)). This refutes the substantive cyclic version of the Erdős–Simonovits compactness claim.

AI involvementAI-led
ImportanceMajor
Human understandingFull

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics
  • Theoretical Computer Science

For the n×n permanent over the complex numbers and n≥32, Chapter 5 gives at least n⁴/(128 log₂ n) variable leaves for division-free arithmetic formulas and n⁴/(192 log₂ n) for valid rational formulas. A separate circuit lower bound has different scope.

AI involvementAI-led
ImportanceNotable
Human understandingFull

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics
  • Theoretical Computer Science

A direct 3SAT reduction proves NP-hardness of approximating Euclidean closest vector within n^(1/400), under the dimension and promise conventions of Chapter 7. The chapter gives separate consequences for binary nearest codeword and finite p norms.

AI involvementAI-led
ImportanceMajor
Human understandingFull

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics
  • Theoretical Computer Science

For the optimal Cohn–Elkies linear-programming packing bound LP_d, OpenAI gives the exponential root limit √(e/(2π)). This yields the sphere-packing upper bound (√(e/(2π))+o(1))ᵈ and corresponding asymptotic Fourier sign-uncertainty radii.

AI involvementAI-led
ImportanceMajor
Human understandingPartial

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics
  • Theoretical Computer Science

OpenAI gives a finitely presented group that lacks finite permutation approximations, establishing the existence of non-sofic groups. Its construction combines a Kun–Thom criterion with an expander-matching argument involving binary Leavitt algebra and Thompson’s group V.

AI involvementAI-led
ImportanceMajor
Human understandingFull

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics
  • Theoretical Computer Science

For every positive dimension n, a compact convex body with nonempty interior, barycenter 0 and no interior lattice point other than 0 has volume at most (n+1)ⁿ/n!. A centered simplex attains equality. This is the inequality, separate from the later equality classification.

AI involvementAI-led
ImportanceMajor
Human understandingPartial

Ten Advances in Mathematics and Theoretical Computer Science

OpenAI

  • Mathematics
  • Theoretical Computer Science

OpenAI constructs infinitely many pairwise nonisomorphic finitely generated ICC property-(T) groups whose tracial group von Neumann factors are mutually isomorphic. The family refutes Connes’s rigidity conjecture and a finite-to-one strengthening.

AI involvementAI-led
ImportanceMajor
Human understandingFull

Signed Depth Relevance for subDL

GPT-5.6 Sol

  • Mathematics
  • Theoretical Computer Science

For every subDL theorem A → B, GPT-5.6 Sol’s terminal-hierarchy argument claims that A and B share a variable with the same sign at the same relevant-arrow depth.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationPartial
Human understandingFew

Online Shadow Tomography Matching the Classical Bounds

Sitan Chen, Ryan O'Donnell, Angelos Pelecanos, John Wright

  • Mathematics
  • Theoretical Computer Science

For an unknown d-dimensional state and m adaptively chosen observables 0≤A≤I, Chen, O’Donnell, Pelecanos and Wright give online expectation estimates within ε with success probability at least 9/10 using O(log(m)√log(d)/ε³) or O(√m/ε²) copies, respectively.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Lean-verified lower bounds for the Shannon capacity of odd cycles

Pjotr Buys, Sven Polak, Jeroen Zuiddam

  • Mathematics
  • Theoretical Computer Science

Buys, Polak and Zuiddam give Lean-formalized Shannon-capacity lower bounds for C7, C11, C13, C15, C19, C21 and C23 using seven explicit base tuples and Gao’s product lemma. Their C7 bound is 3.258805369885… and their C11 bound is 5.294502522149….

AI involvementAI-centered
ImportanceModest
Technical verificationFormal (Lean + human)
Human understandingFull

On the Integer Domination Root Conjecture

Saeid Alikhani, Max Griswold

  • Mathematics
  • Theoretical Computer Science

A specified connected graph with 33 vertices and 36 edges has domination polynomial divisible by x+4. Its exact factorization includes x¹¹(x+2)(x+4)(x²+3x+1), so −4 is an integer domination root.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

On a conjecture of Han and Xiong for fractional Gaussian binomial coefficients

Ken Ono

  • Mathematics
  • Theoretical Computer Science

For positive rational u=c/d and r=a/b in lowest terms, with d even, d dividing lcm(b,2), and u≤r, the integer trace of the fractional Gaussian binomial coefficient at u dominates that at r coefficientwise for every k≥1. This proves the Han–Xiong comparison for every r≥1/2 and reduces the remaining parameters to 1/(2m).

AI involvementAI-autonomous
ImportanceNotable
Human understandingFull

From 12 to 6: Sharpening the Three-Charge Bound in Maxwell's Problem

A. Gabrielov, Dm. Novikov, T. Novikov, B. Shapiro

  • Mathematics
  • Theoretical Physics

For every α>0 and three positive point charges in R^n with n≥2, the potential has at most six nondegenerate equilibria. The new step proves a four-contact statement using separation at the unique saddle of a first integral.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Local cohomology modules with nonclosed support

Bhargav Bhatt, Linquan Ma

  • Mathematics

Bhatt and Ma construct a five-dimensional finitely generated flat normal domain over Z or any field with a codimension-one prime J for which H²_J has infinitely many minimal primes. The characteristic-two case uses a separate construction; a cone gives a six-dimensional normal local example.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Settling the Optimal Exponent Relating Sumsets and Difference Sets

Haowei Lin, Shanda Li

  • Mathematics

For finite integer sets A with at least two elements, Lin and Li construct families with log(|A+A|/|A|)/log(|A−A|/|A|) approaching 2. Together with Ruzsa’s upper bound, this proves that the exponent 1/2 in √σ(A)≤δ(A) is sharp.

The Maxwell Conjecture is False

Philip Arathoon, Gavin Ball, Matthew D. Kvalheim

  • Mathematics
  • Theoretical Physics

Five positive point charges in R³ can have at least 24 nondegenerate electrostatic equilibria. A small perturbation makes all critical points nondegenerate and finite in number, exceeding Maxwell’s proposed (n−1)² bound of 16 when n=5.

AI involvementAI-centered
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Removing the Torsion-free Hypothesis in a Positivity Theorem on Deligne-Mumford Stacks

Shengyu Hou, Jihao Liu

  • Mathematics

Hou and Liu prove pseudo-effectivity of the first Chern class for every coherent quotient of every positive tensor power of the logarithmic cotangent sheaf under their smooth proper integral complex Deligne–Mumford stack, normal-crossings divisor, and big-determinant subsheaf hypotheses.

AI involvementAI-centered
ImportanceModest
Technical verificationFormal (human)
Human understandingFull

Bosonic quantum communication beyond the thermal threshold

Francesco Anna Mele, Giuseppe Catalano, Marco Fanizza, Vittorio Giovannetti, Ludovico Lami

  • Mathematics
  • Theoretical Physics

For a thermal attenuator with environment photon number ν=1, a rank-two input on six Fock levels has certified coherent information at least 4.7×10⁻⁴ qubits per channel use at transmissivity η=0.8, while all single-mode Gaussian inputs have at most zero. A larger witness is certified positive at η=0.7841.

Optimal partial plank coverings

Egor Bakaev, Alexander Polyanskii

  • Mathematics

For any planar convex body K and 0<W<width(K), one plank of width W covers at least as much area of K as any finite family of planks whose total width is at most W. The assessed AI-associated result is this planar reduction.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Using large language models to probe the limits of atom-centered structural descriptors

Michelangelo Domina, Michele Ceriotti

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

Domina and Ceriotti give noncongruent atom-centered environments with identical complete correlations through seven-neighbor order, including three-dimensional lifts. For every finite correlation order and angular cutoff, they also construct indistinguishable noncongruent environments.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

Tournaments determined by three and five voters

Leonid Chindelevitch, Ararat Harutyunyan

  • Mathematics
  • Theoretical Computer Science

The Paley tournament on 43 vertices has predictability 181/301, above the five-voter threshold 3/5, yet cannot be induced by five voters. The proof uses a co-backing constraint for cyclic triangles and an exact finite shell-and-screen enumeration.

A Proposed Solution to Erdős Problem 390

Shouqiao Wang

  • Mathematics
  • Theoretical Computer Science

For the least possible largest distinct factor f(n) in n! = a₁⋯aₖ with n<a₁<⋯<aₖ, Wang’s manuscript claims that (f(n)−2n)log(n)/n tends to 4029639598/25970038185. Its new step is a matching upper-bound construction.

AI involvementAI-led
ImportanceNotable
Human understandingFew

Turbulent Flame Speed Can Increase under Curvature Smoothing

Hung V. Tran, Jack Xin, Yifeng Yu

  • Mathematics

Tran, Xin and Yu construct a smooth periodic three-dimensional shear flow V=(0,0,f(x₁,x₂)) and direction (2,2,1) whose homogenized G-equation flame speed has c(0)=0 and positive right derivative as curvature diffusivity d tends to zero. For small d the cutoff is inactive, so the physical cutoff cell speed increases.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Sharp bounds for the fractional chromatic number of high-girth d-degenerate graphs

Peter Allen, Abhishek Dhawan, Jonathan A. Noel

  • Mathematics
  • Theoretical Computer Science

Allen, Dhawan and Noel prove that for every fixed girth g≥4 there are d-degenerate graphs of girth at least g with fractional chromatic number at least (1−o(1))d/log d as d grows. ChatGPT solved the authors’ optimization problem for the fractional clique function used in this lower-bound proof.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Kazhdan-Lusztig polynomials of matroids need not be unimodal

Ronnie Cheng, Shurui Liu

  • Mathematics

Rethlas found a representable rank-seven binary projective-deletion matroid whose Kazhdan–Lusztig polynomial is not real-rooted, with a convolution and degree approach. Cheng and Liu later used that direction to construct representable matroids over every finite field with nonunimodal Kazhdan–Lusztig polynomials.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

A Negative Answer to Stanley's Problem 4 on Differential Posets

Xinan Dai, Wenhao Deng, Yingdong Shi, Tailin Wu, Yuchen Yang

  • Mathematics

Dai and coauthors construct a locally finite 1-differential poset P for which the two-multichain quotient M_(P,2)(q)/F_P(q)² is nonrational over every characteristic-zero field. This refutes Stanley’s proposed rational-factor property already at k=2.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

The exact solution of Bellman's lost-in-a-forest problem for the golden gnomon

Alexander Temerev, Alessio Doria

  • Mathematics
  • Theoretical Computer Science

Temerev and Doria report that the minimum escape-path length for the golden gnomon, an isosceles triangle with unit equal sides and 36-degree base angles, is 1.282676025459… . A symmetric seven-piece path attains the value, and their lower bound covers every rectifiable competitor.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFew

Sharp continuity of quantum conditional entropy

Mario Berta, Pablo Costa Rico, Gereon Kossmann, Ludovico Lami, Julius A. Zeiss

  • Mathematics
  • Theoretical Physics

For finite-dimensional quantum states at trace distance at most δ and d=dim A, the optimal dimension-only conditional-entropy continuity modulus is h₂(δ)+δ log(d²−1) up to δ=1−d⁻², then 2 log d. Equality is attainable when dim B≥d.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Sharp small-deviation inequalities for sums of independent nonnegative random variables

Weibo Fu, Yanjun Han, Guanyang Wang, Jun Yan, Peng Zhang, Zhengqing Zhou

  • Mathematics

For independent nonnegative random variables X₁,…,Xₙ with E[Xᵢ]≤1 and S=ΣXᵢ, Fu, Han, Wang, Yan, Zhang and Zhou prove P(S<E[S]+δ)≥δ(n/(n+δ))ⁿ for 0<δ<1, and ≥(1−1/(n+δ))ⁿ for δ≥1. The second finite-n bound is sharp and implies Feige’s unit-slack e⁻¹ conjecture.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

A proof of the strong Gaussian product inequality conjecture

Frédéric Ouimet, Dylan Greaves

  • Mathematics
  • Theoretical Computer Science

For any centered Gaussian vector, including singular covariance, and positive real exponents αᵢ, the expected product of |Xᵢ|^αᵢ is at least the product of the corresponding marginal expectations.

AI involvementAI-led
ImportanceNotable
Human understandingFull

AI-led work gives an explicit 2-adic absolute Galois presentation

  • Mathematics
  • Theoretical Computer Science

Theorem 1.2 gives a marked profinite presentation of the absolute Galois group of Q₂ with four generators, a tame relation, an explicit wild-word relation, and a pro-2 condition on the normal subgroup generated by the wild generators.

AI involvementAI-led
ImportanceMajor
Human understandingFull

Powers of the Vandermonde determinant are eventually non-SNP

Thien Le, Melanie Weber

  • Mathematics
  • Theoretical Computer Science

For every fixed power k≥1, sufficiently many variables give a lattice monomial in the Newton polytope of the k-th Vandermonde power with zero coefficient. The new step is an explicit even-power seed for k≥4; earlier cases and propagation complete the result.

A partial-trace matrix inequality and Werner-state distillability

Zhiwei Song, Lin Chen

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

GPT-5.5 Pro’s Proposition 2 expresses the rank-two partial-trace quadratic form q⁽²⁾₋₁/₂(C) as projection terms. Song and Chen then use a separate human-developed lemma to prove nonnegativity and two-copy nondistillability at the Werner threshold.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

An Exact Counterexample to Carlson's Associated-Prime Depth Conjecture from a Group of Order 128

Xinan Dai, Wenhao Deng, Yingdong Shi, Tailin Wu, Yuchen Yang

  • Mathematics

For G=SmallGroup(128,859) over the algebraic closure of F₂, depth H*(G;k)=2 while every associated-prime quotient has dimension at least 3. The group-specific certificate enumerates 75 rank-two elementary abelian subgroups and checks their centralizer types.

An Explicit Counterexample to Stanley's Rankwise Lower-Bound Conjecture for Differential Posets

Xinan Dai, Wenhao Deng, Yingdong Shi, Tailin Wu, Yuchen Yang

  • Mathematics

For each integer r≥3, Dai and coauthors construct an infinite r-differential poset that agrees with Yʳ through rank three but has rank-four size |(Yʳ)₄|−⌊r/3⌋. At r=3 the first sizes are 1, 3, 9, 22, 50, compared with 51 for Y³.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Exponentially Fewer-Server PIR from Sparser S-Decoding Polynomials

Aparna Gupte, Seyoon Ragavan

  • Mathematics
  • Theoretical Computer Science

Conditional on Conjecture 1.1 about large prime divisors, the authors construct an S*_m-decoding polynomial with the optimal k+1 nonzero terms for every fixed k, for suitable k-prime m and finite field. The construction gives the stated fixed-server PIR tradeoff; finite prime checks make it unconditional through 15 servers.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Matroid flat counts are not unimodal

Alexander Divoux, Chayim Lowen, Shouda Wang

  • Mathematics

The authors adapt a ChatGPT-supplied q-lift example against reciprocal flat-count convexity into matroids whose numbers of flats by rank are not unimodal. The final construction combines Larson’s graphic matroids with Whittle’s q-lift.

AI involvementAI-centered
ImportanceMajor
Technical verificationReported
Human understandingFull

Commutator relators of one-relator groups do not force Hopficity, residual finiteness, or automaticity

Ke Wang, Qiang Zhang

  • Mathematics

For G_m=⟨a,t | [t,a[a,t]^(−m)]⟩, a Baumslag–Solitar group BS(m,m+1) is a retract. In particular, G_2 is non-Hopfian and non-residually finite, while G_1 is non-automatic. Literal commutator relators therefore do not force any of these properties.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Chromatic symmetric functions of claw-free graphs are not Schur positive

Jacob P. Matherne, Alejandro H. Morales

  • Mathematics

Matherne and Morales give a 12-vertex, 19-edge bipartite graph G3 whose chromatic symmetric function is reported Schur positive while its three-variable specialization omits lattice-point weight (7,4,1) between present weights (6,6,0) and (8,2,2).

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

Unconditional Unclonable Encryption

Prabhanjan Ananth, Seyoon Ragavan, Amit Sahai

  • Mathematics
  • Theoretical Computer Science

Ananth, Ragavan and Sahai give efficient, perfectly correct, one-time private-key encryption of one bit using a random nonidentity phase-free n-qubit Pauli and a random eigenstate. After key disclosure, both separated receivers recover the bit with probability at most 1/2+O(2^(−n/2)) against information-theoretic attacks.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4 + human)
Human understandingFull

Chromatic symmetric functions of claw-free graphs are not Schur positive

Jacob P. Matherne, Alejandro H. Morales

  • Mathematics

Matherne and Morales give two 12-vertex claw-free line graphs whose chromatic symmetric functions have [s₃₃₃₃] coefficients −64 and −40. Either graph refutes Stanley’s claim that every claw-free graph is Schur-positive.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

An Explicit Counterexample to the Rank-Two Poisson Conjecture

Christopher D. Long

  • Mathematics

Long gives explicit rational-coefficient polynomials in four variables defining a Poisson endomorphism of the canonical algebra P₂ in characteristic zero that is not an automorphism. An exact three-point fiber proves noninjectivity; extending the map by identity gives examples for every Pₙ with n≥2.

AI involvementAI-centered
ImportanceMajor
Technical verificationPartial
Human understandingFew

An Improved Upper Bound for Colorings Without Symmetrically Colored k-Term Arithmetic Progressions

Ruizhe Shi, Yiqi Dong

  • Mathematics

For every even k≥4 and prime p>k, a coloring of Z/(p^(k²/4))Z uses (k−1)^(k²/4−1)p colors and avoids nontrivial symmetrically colored k-term arithmetic progressions. At k=4 this gives 27p colors and an O(N^(1/4)) interval bound.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Polynomial Hilbert-Schmidt stability of the lamplighter group

Alon Dogon, Thomas Vidick

  • Mathematics
  • Theoretical Computer Science

For finite-dimensional unitaries satisfying approximate order-two and orbit-commutator relations through 2M, with M polynomial in the target tolerance, Dogon and Vidick prove nearby exact lamplighter relations. Their argument gives polynomial upper bounds on Hilbert–Schmidt stability rate and radius.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Dittert's conjecture in dimension 16 via a joint-deficit scaling lemma

Boris Kafidov

  • Mathematics

Kafidov claims that, for every nonnegative 16-by-16 matrix with total entry sum 16, the Dittert functional (product of row sums plus product of column sums minus permanent) is at most 2-16!/16^16, with equality only at the constant matrix. The new joint-deficit scaling lemma makes a near maximizer doubly superstochastic after a sharper dilation and excludes a boundary maximizer at dimension 16.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Chernoff's Density Is Strongly Log-Concave

Xianyang Zhang, Quan Zhou

  • Mathematics

Theorem 1 proves that the one-sided Airy factor g has v(x) = -(log g)''(x) > 0 and v(x) ≥ 4x. Since Chernoff's density f is proportional to g(x)g(-x), its curvature κ(x) exceeds 4|x| and has a positive global infimum. This proves strong log-concavity of the density of argmax_t{W(t)-t²}.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (human)
Human understandingFew

Counterexamples to the xz-Conjecture and the Mathieu Conjecture for SU(2)

Christopher D. Long

  • Mathematics
  • Theoretical Computer Science

The Laurent polynomial f(x,z)=(1−z⁻¹)((1−x)+xz) has zero pure product-integral moments for every positive n but marked moment (−1)^(n−1)/(n+1). It refutes xz(1,1), with padding for all mixed k,l≥1. Its SU(2) lift F=(1+c)(ad+b), G=−c has the same pure and marked Haar moments and refutes the SU(2) Mathieu conjecture.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Completeness of Canonical Closure Representations Is coNP-Complete

Mikhail Babin

  • Mathematics
  • Theoretical Computer Science

Testing whether a supplied family of meet-irreducibles is the complete canonical family of a finite implication closure system is coNP-complete, even with acyclic implications of premise size at most three and a genuine irredundant supplied family. The hard direction reduces 3-UNSAT to a missing member.

AI involvementAI-led
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Small Counterexamples to the Gaussian Moments Conjecture

Christopher D. Long

  • Mathematics

Christopher D. Long reports explicit complex polynomials P and Q in three independent real Gaussian variables with E(P^m)=0 and E(QP^m)=m! for every positive integer m. A five-term degree-four example refutes the conjecture in every dimension at least three; a six-term cubic example works in four variables.

AI involvementAI-autonomous
ImportanceMajor
Technical verificationPartial
Human understandingFull

A Globally Asymptotically Stable Planar Homogeneous Polynomial Vector Field With No Polynomial Lyapunov Function

Jun Liu, Maxwell Fitzsimmons

  • Mathematics
  • Theoretical Computer Science

Liu and Fitzsimmons give an explicit planar homogeneous cubic polynomial vector field with integer coefficients whose origin is globally asymptotically stable and has a positive radially unbounded C¹ homogeneous Lyapunov function, but has no positive definite homogeneous polynomial with nonpositive Lie derivative and no local real-analytic weak Lyapunov function.

Semi-Streaming Matching in a Single Pass II: Greedy is Optimal

Sepehr Assadi, Max Jiang, Mars Xiang

  • Mathematics
  • Theoretical Computer Science

Assadi, Jiang and Xiang construct optimal blueprints showing that no deterministic or randomized single-pass semi-streaming algorithm using O(n polylog n) memory and constant success probability can guarantee better than a one-half approximation for maximum bipartite matching. Greedy attains one half.

AI involvementAI-centered
ImportanceMajor
Technical verificationPartial
Human understandingFull

The analogue of Belinskaya's theorem for measure-preserving flows

Konstantin Slutsky

  • Mathematics

A measurable real-line set with finite average symmetric-difference measure under unit-interval translations is commensurate with exactly one of the empty set, the whole line, a right half-line or a left half-line. For free ergodic measure-preserving flows, this implies that abstractly isomorphic L¹ full groups determine conjugacy up to a nonzero scalar time change.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Online Beck--Fiala Down to Logarithmic Sparsity

Dylan J. Altschuler, Konstantin Tikhomirov

  • Mathematics

For every fixed sequence of T=n vectors in [−1,1]^n chosen before the algorithm runs, the model-generated proof gives O(√n) maximum online prefix discrepancy. The authors also prove longer-horizon and sparse-column extensions, including O(√n) for T up to exp(cn/log^(2+η)(en)) with the stated high probability.

AI involvementAI-centered
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Star Fleet reports a counterexample to Erdős Problem 769’s lower bound

  • Mathematics
  • Theoretical Computer Science

Star Fleet's construction shows that for every odd n at least 201, every integer k at least n·2^n·ceil(49n/100)^n+2 is the count of a tiling of the unit n-cube by positive homothetic cubes. This gives c(n)=o(n^n) along odd dimensions and refutes the proposed universal lower bound c(n)≫n^n; it does not determine the best bounds for c(n) in all dimensions.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

A counterexample to a conjecture of Thakur on Carlitz-Wieferich primes

David Niedbala Giraudin

  • Mathematics

Niedbala Giraudin gives an explicit irreducible degree-five Carlitz–Wieferich prime over F_(19³), refuting the proposed odd-characteristic rule that its degree must be divisible by the characteristic.

AI involvementAI-led
ImportanceNotable
Human understandingPartial

Record compositions of alternating permutations and noncommutative symmetric functions

Evan Chen, Ken Ono, Michal Mogielnicki

  • Mathematics
  • Theoretical Computer Science

For each composition α=(α₁,…,α_k) of n, the number of alternating permutations of 2n with record composition α is ∏_j binom(2s_j−1,2α_j−1)E_(2α_j−1), where s_j=α₁+…+α_j and E_m is an Euler number. The paper also realizes these counts in a noncommutative symmetric-function lift.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4.31.0)
Human understandingFull

Graph Puzzles III.1: A Proof of Sabidussi's Compatibility Conjecture

Nikolay Ulyanov

  • Mathematics
  • Theoretical Computer Science

Every finite connected Eulerian multigraph of minimum degree at least four has, for a fixed Euler tour, an edge partition into connected 2-regular circuits compatible with that tour’s transitions.

AI involvementAI-led
ImportanceNotable
Human understandingFull

A dual linear programming bound for sphere packing in dimension 36

Rifat Jumagulov

  • Mathematics
  • Theoretical Computer Science

Jumagulov certifies that the dimension-36 Cohn–Elkies two-point LP optimum is at least 146.1036734821, more than 32.91 times the Kschischang–Pasupathy packing’s center density.

The storage capacity of the Ising perceptron: verification of the outstanding numerical conditions

Yitzchak Shmalo

  • Mathematics
  • Theoretical Physics

Shmalo's paper verifies Ding–Sun Condition 1.2, including curvature and endpoint clauses, and Huang Condition 1.3 for the zero-margin Ising perceptron. Combining these with prior conditional and universality theorems yields M_N/N→α★≈0.8330786 in probability for Gaussian and fixed standardized subgaussian disorder, including Bernoulli.

AI involvementAI-autonomous
ImportanceMajor
Technical verificationPartial
Human understandingFew

The Benjamini--Hochberg Procedure Can Fail to Control the FDR for Correlated Two-Sided Gaussian Tests

Edgar Dobriban

  • Mathematics

In a specified three-block Gaussian factor model with 96N true nulls and 4N nonnulls, ordinary Benjamini–Hochberg testing at level 0.01 has false discovery rate greater than 0.0104 for all sufficiently large N when applied to correlated two-sided Gaussian p-values.

AI involvementAI-autonomous
ImportanceNotable
Human understandingFull

The perfect 1-factorisation conjecture holds asymptotically

Yangyang Cheng, Amedeo Sgueglia

  • Mathematics
  • Theoretical Computer Science

For every ε>0 and all sufficiently large even n, the complete graph K_n contains at least (1−ε)n edge-disjoint perfect matchings such that every pair forms a Hamilton cycle. This is the model-origin construction of Theorem 1.3.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Near-optimal node-private community estimation in polynomial-time

Laurentiu Marchis, Olga Klopp, Po-Ling Loh, Ilias Zadik

  • Mathematics
  • Theoretical Computer Science

Under the stated balanced stochastic-block-model signal and community-growth assumptions, the authors give a high-probability polynomial-time pure node-private estimator attaining the prior exponential-time method’s minimax exact-recovery risk, with privacy parameter of order log n.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Cantor measures with odd base do not admit Fourier frames

Jaume de Dios Pont, Lukas Liehr, Mitchell A. Taylor

  • Mathematics
  • Theoretical Computer Science

For every odd integer base b>1, the fair {0,2} base-b Cantor measure admits no Fourier frame. The middle-third case answers Strichartz’s question.

AI-found dumbbell graph refutes Graffiti Conjecture 143

  • Mathematics
  • Theoretical Computer Science

A 39-vertex dumbbell graph violates Graffiti Conjecture 143, which bounds the variance of positive adjacency eigenvalues by graph size divided by average distance. The witness works under both standard average-distance conventions.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFew

AI-found values refute both sign clauses of Sun’s Conjecture 4.6(ii)

  • Mathematics
  • Theoretical Computer Science

For the integer-normalized trigonometric permanents at prime p=29, s₂₉=1,053,859 is positive and s′₂₉=−4,806,838,304 is negative. These exact values contradict both sign predictions in Conjecture 4.6(ii).

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFew

AI-found graphs refute Graffiti 154 under a standard-deviation reading

  • Mathematics
  • Theoretical Computer Science

Lollipop graphs with 118 or 120 vertices violate Graffiti Conjecture 154 when “deviation” means the standard deviation of adjacency eigenvalues. A 144-vertex graph violates it under both standard average-distance conventions.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFew

A proof of Kurkov's conjecture for A000670

John Erlbacher

  • Mathematics

For every n>0, the Fubini number a(n), which counts ordered set partitions, equals the sum of A284005(k) for k from 0 through 2^(n−1)−1. The proposed proof counts partitions with each prescribed set of block minima.

AI involvementAI-centered
ImportanceModest
Technical verificationPartial
Human understandingFew

Involution-equivariant topological recursion and mirror symmetry for the affine binary dihedral Calabi--Yau threefold

Bohan Fang, Zhuoming Lan, Jingxiang Ma

  • Mathematics
  • Theoretical Physics

Fang, Lan and Ma identify affine binary-dihedral orbifold Gromov–Witten theory with involution-equivariant sign-sector recursion on a type-D logarithmic Toda curve, for stable descendants with n>0 and 2g−2+n>0 and free energies with g≥2 on the stated smooth chamber.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Stable Phase Retrieval for Spans of Independent Random Variables

Pedro Abdalla, Jaume de Dios Pont, João P. G. Ramos, Mitchell A. Taylor

  • Mathematics
  • Theoretical Computer Science

Theorem 6.1 gives an explicit quantitative stable-phase-retrieval bound for the closed real L² span of independent centred normalized random variables under indexed two-sided L¹ bounds on all but at most one coordinate. Its proof uses a head–tail concentration versus diffusion argument and Sperner-type anticoncentration.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

Integer values of tan(arctan 1 + arctan 2 + ⋯ + arctan n) are rare

Ken Ono

  • Mathematics
  • Theoretical Computer Science

AxiomProver proved in Lean that an integral x_n forces the squarefree kernel K_n to divide 1+m² with |m|≥√(K_n−1), that large exceptional x_n lie near a π/2 lattice point, and that the exceptional set has O(log N) members through N. Ono combines these with his asymptotic lower bound for K_n to show nonintegrality for almost all n.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4.28.0 / Mathlib)
Human understandingFull

The singleton hypergraph is extremal for the Isolation Lemma

Vance Faber, David G. Harris

  • Mathematics
  • Theoretical Computer Science

For every nonempty inclusion-free hypergraph on n vertices, every d≥1 and arbitrary real edge offsets, Faber and Harris prove that at least n times the sum of j^(n−1) for j=0,…,d−1 assignments in [d]^n isolate a unique minimum edge. The n-singleton hypergraph attains equality.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Deletion-Star Ceilings and a 1.28249… Lower Bound for Cost-Preserving Single-Source Unsplittable Flows

Sergey Nikolenko

  • Mathematics
  • Theoretical Computer Science

Nikolenko constructs a 17-terminal common-point instance proving the universal cost-preserving unsplittable-flow constant is at least 1.282494797984843521. A 3/2 congestion ceiling applies only to common-point cells of complement mass two.

AI involvementAI-led
ImportanceNotable
Technical verificationPartial
Human understandingFull

Exact-Mean Decompositions for Cost-Preserving Single-Source Unsplittable Flows: Demand Scales and Local Interaction Structure

Sergey Nikolenko

  • Mathematics
  • Theoretical Computer Science

Nikolenko gives a six-terminal planar instance requiring congestion factor at least 58676765987259/50000000000000, or 1.17353531974518, under cost nonincrease. He also identifies exact local envelopes E(2)=1, E(3)=9/8 and E(4)=(299−41√41)/32.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

A Counterexample to Ziegler's Cross-Polytope Conjecture for Simplicial 0/1-Polytopes

Volker Kaibel, Sebastian Pokutta

  • Mathematics

Fourteen binary vertices form a simplicial seven-dimensional 0/1-polytope with the maximum 2d vertices but without central symmetry, countering the proposed cross-polytope characterization.

Criteria of isolated weighted homogeneous hypersurface singularities using Logarithmic vector fields

Jihao Liu, Xiping Zhang

  • Mathematics

For a reduced isolated complex hypersurface germ of dimension at least two, or an irreducible curve germ, weighted homogeneity after coordinate change is equivalent to a logarithmic holomorphic vector field transverse in the real-Euclidean sense to every sufficiently small link in those same coordinates. An invertible ambient linear part is a third equivalent condition.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationPartial
Human understandingFull

Rectangles, triangles and Schrödinger waves

Jonathan Bennett, Vjekoslav Kovač, Shohei Nakamura, Itamar Oliveira

  • Mathematics

Finite planar lattice sets can have an arbitrarily large ratio of ordered rectangles to ordered isosceles triangles, including degenerate configurations. Bennett, Kovač, Nakamura and Oliveira quantify the construction in a box [-N,N]² by a ratio at least (log N)^(c/log log log N), with an analogous cardinality bound.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Furthest Pair Requires Quadratic Time in Superconstant Dimension under SETH

Barna Saha, Yinzhan Xu, Christopher Ye

  • Mathematics
  • Theoretical Computer Science

Under OVH or SETH, Saha, Xu and Ye prove n^(2−o(1)) lower bounds for exact Furthest Pair, Bichromatic Closest Pair, Maximum Inner Product and Hopcroft’s problem in every efficiently constructible superconstant dimension.

AI involvementAI-led
ImportanceMajor
Human understandingFull

Counterexample to a conjecture on the pairwise independent correlation gap using AI

Arjun Ramachandra, Karthik Natarajan

  • Mathematics
  • Theoretical Computer Science

A five-element coverage function with marginals (3/10, 7/20, 3/10, 7/20, 7/20) has ordinary concave closure 4 and pairwise-independent closure at most 479/160. Their ratio is at least 640/479, above the conjectured universal 4/3 bound.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Counterexamples to Wegner's Conjecture for Rectangles

Deepak Ajwani, Rishikesh Gajjala, Rajiv Raman, Saurabh Ray

  • Mathematics

Ajwani, Gajjala, Raman and Ray give 64 axis-parallel rectangles with packing number 16 and piercing number at least 32, disproving Wegner’s proposed bound 2ν−1. Their package construction yields piercing and clique-LP ratios approaching 5/2.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

A Resolution of the SS--RS--GD Inequalities

Binghui Peng

  • Mathematics
  • Theoretical Computer Science

For n=3, K=2 and d=4, Peng gives positive definite symmetric matrices arbitrarily close to the identity with ||W_SS|| > ||W_RS||, refuting the SS–RS half. For all n≥2, K≥1 and d≥1, the RS–GD half holds when (1−1/(4n²+1))I ≤ Aᵢ ≤ I.

AI involvementAI-led
ImportanceNotable
Human understandingFull

On Injectivity of Phase Retrieval

Zhangsong Li

  • Mathematics
  • Theoretical Computer Science

For every M≥2 with N=4M−5, a nonempty open set of complex N×M measurement matrices has a noninjective phase-retrieval map. Hence an IID complex Gaussian matrix is noninjective with positive probability, so p_M<1.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Claude counterexample refutes Erdős #619 diameter-four saving

  • Mathematics

For every η>0, a family of connected triangle-free graphs needs at least (1−η)n new edges to become diameter at most four while staying triangle-free. The problem record states the stronger quantitative bound h₄(G)≥n−O(n^(8/9)(log n)^(2/9)).

AI involvementAI-led
ImportanceNotable
Human understandingFew

A proof of an identity for the critical exponents of jamming

Giorgio Parisi, Francesco Zamponi

  • Mathematics
  • Theoretical Physics

Parisi and Zamponi give an analytic proof of a+b=1 from the scaling full replica-symmetry-breaking equations for infinite-dimensional hard-sphere jamming. Together with the earlier relation b=(1+c)/2, it yields the scaling identities linking force, gap and overlap exponents.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

On Thin Perfect Matchings up to Polylogarithmic Factors

Alireza Haqi, Shayan Oveis Gharan

  • Mathematics
  • Theoretical Computer Science

A connection from relaxed thin perfect matching to a single-tree cut sparsifier helps yield an O(log n log log n)-thin perfect matching relative to any fractional perfect matching, allowing edges outside its support. The separate O(n log n) bound requires support-constrained edges.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Sampling Directed Eulerian Tours in Õ(m³ᐟ²) Time

Nima Anari

  • Mathematics
  • Theoretical Computer Science

Anari’s Theorem 4 reports that the flip–repair walk associated with any real skew-symmetric n×n matrix has spectral gap at least 2/n. In the flat 0/1 determinant case it also gives an entropy bound and nearly linear mixing; this feeds a near-uniform sampler for strongly connected directed Eulerian multigraphs with m arcs in O(m^(3/2) polylog(m/ε)) time.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFew

Resolving Thinness of Degree-Six Symplectic Hypergeometric Monodromy Groups

J. Maxwell Riestenberg, Diaaeldin Taha, Steve Trettel, Anna Wienhard

  • Mathematics

AlphaEvolve found explicit witness words that Riestenberg, Taha, Trettel and Wienhard use to prove that the degree-six symplectic hypergeometric monodromy groups C-47 and C-55 have finite index in their ambient integral symplectic groups and are arithmetic.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Intervals of hypergraph Turán densities

Xizhi Liu, Oleg Pikhurko

  • Mathematics

Liu and Pikhurko use a Ramanujan-graph spectral 3-graph gadget to build a smoothly decreasing sequence of Lagrangians. With their separate recursive and lifting constructions, they show that for every r≥3, Turán densities of possibly infinite forbidden r-graph families contain a terminal interval [1−δ_r,1].

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Advancing Mathematics Research with AI-Driven Formal Proof Search

George Tsoukalas, Anton Kovsharov, Sergey Shirobokov, Anja Surina, Moritz Firsching, Gergely Bérczi, Francisco J. R. Ruiz, Arun Suggala, Adam Zsolt Wagner, Eric Wieser, Lei Yu, Aja Huang, Miklós Z. Horváth, Andrew Ferraiuolo, Henryk Michalewski, Edward Lockhart, Codrut Grosu, Thomas Hubert, Matej Balog, Pushmeet Kohli, Swarat Chaudhuri

  • Mathematics
  • Theoretical Computer Science
  • Theoretical Physics

The AlphaProof Nexus report says its agent ruled out monochromatic quantum graphs for the diagonal four-, six- and ten-particle cases N=d=4, 6 and 10 in the Krenn–Gu perfect-matching model.

AI involvementAI-led
ImportanceNotable
Technical verificationReported
Human understandingFew

Advancing Mathematics Research with AI-Driven Formal Proof Search

George Tsoukalas, Anton Kovsharov, Sergey Shirobokov, Anja Surina, Moritz Firsching, Gergely Bérczi, Francisco J. R. Ruiz, Arun Suggala, Adam Zsolt Wagner, Eric Wieser, Lei Yu, Aja Huang, Miklós Z. Horváth, Andrew Ferraiuolo, Henryk Michalewski, Edward Lockhart, Codrut Grosu, Thomas Hubert, Matej Balog, Pushmeet Kohli, Swarat Chaudhuri

  • Mathematics
  • Theoretical Computer Science

For every finite simple connected graph G, the maximum number L_S(G) of leaves in a spanning tree is at least 2(ℓ(G)−1), where ℓ(G) is the average local independence number.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

Advancing Mathematics Research with AI-Driven Formal Proof Search

George Tsoukalas, Anton Kovsharov, Sergey Shirobokov, Anja Surina, Moritz Firsching, Gergely Bérczi, Francisco J. R. Ruiz, Arun Suggala, Adam Zsolt Wagner, Eric Wieser, Lei Yu, Aja Huang, Miklós Z. Horváth, Andrew Ferraiuolo, Henryk Michalewski, Edward Lockhart, Codrut Grosu, Thomas Hubert, Matej Balog, Pushmeet Kohli, Swarat Chaudhuri

  • Mathematics
  • Theoretical Computer Science

For every n-element Sidon set A of natural numbers, AlphaProof Nexus proves 16 I(A+A)+100n+16≥n², where I counts sums whose neighboring integers are both absent from A+A. Thus the minimum isolated-sum count grows at least quadratically, answering the finite-set form of Erdős Problem 152.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

Advancing Mathematics Research with AI-Driven Formal Proof Search

George Tsoukalas, Anton Kovsharov, Sergey Shirobokov, Anja Surina, Moritz Firsching, Gergely Bérczi, Francisco J. R. Ruiz, Arun Suggala, Adam Zsolt Wagner, Eric Wieser, Lei Yu, Aja Huang, Miklós Z. Horváth, Andrew Ferraiuolo, Henryk Michalewski, Edward Lockhart, Codrut Grosu, Thomas Hubert, Matej Balog, Pushmeet Kohli, Swarat Chaudhuri

  • Mathematics

Every pure O-sequence of codimension three and type two is log-concave: h(i−1)h(i+1)≤h(i)^2 at every interior degree.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

Advancing Mathematics Research with AI-Driven Formal Proof Search

George Tsoukalas, Anton Kovsharov, Sergey Shirobokov, Anja Surina, Moritz Firsching, Gergely Bérczi, Francisco J. R. Ruiz, Arun Suggala, Adam Zsolt Wagner, Eric Wieser, Lei Yu, Aja Huang, Miklós Z. Horváth, Andrew Ferraiuolo, Henryk Michalewski, Edward Lockhart, Codrut Grosu, Thomas Hubert, Matej Balog, Pushmeet Kohli, Swarat Chaudhuri

  • Mathematics
  • Theoretical Computer Science

A counterexample on Z/3Z disproves the intended complex-valued equality of two quadratically structured function spaces in Ben Green’s Problem 57.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFew

Optimal e^((γ+o(1))n)-Approximation of the Permanent of Positive Semidefinite Matrices

Nima Anari, Farzam Ebrahimnejad

  • Mathematics
  • Theoretical Computer Science

For an n×n Hermitian positive semidefinite matrix A, an entropy-corrected concave Gram-factor relaxation P̂ satisfies e^(−γn)P̂≤per(A)≤P̂ and gives a deterministic polynomial-time e^((γ+ε)n)-approximation for every ε>0. The exponent is optimal up to lower-order terms under the cited NP-hardness, assuming P≠NP.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Shokurov's global index conjecture for threefold foliations

Jihao Liu, Sheng Qin

  • Mathematics

Liu and Qin prove a uniform index I for projective log-canonical foliated triples of dimension at most three with K_F+B numerically trivial and coefficients in a fixed rational DCC set, so I(K_F+B) is linearly trivial. Rethlas contributed the non-algebraically-integrable case split and Lie-theoretic route.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

On a question of Mauri and Moraga

Jihao Liu

  • Mathematics

A surface qdlt log Calabi–Yau pair with three big boundary components has a square dual complex, rather than the boundary of a two-dimensional standard simplex. This answers part (2) of Mauri and Moraga’s Question 9.9 negatively.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationReported
Human understandingFew

Reciprocals of Partition Polynomials

Evan Chen, Ken Ono, Jujian Zhang

  • Mathematics
  • Theoretical Computer Science

AxiomProver generated Lean proofs of Ballantine–Beck–Feigon–Maurischat Conjectures 2, 5, 7, 8, 9 and the intended corrected reading of 10, covering coprimality, divisibility, special values and a recurrence. It also found an n=4 counterexample to a printed binary log-concavity claim.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

On Talagrand's Convexity Conjecture

Dongming Merrick Hua, Antoine Song, Stefan Tudose

  • Mathematics

Appendix B Proposition B.1 gives a convex polyhedral Gaussian partition with prescribed cell masses and projected barycenters when a centered finite-support X is convexly dominated by ρG for 0<ρ<1. The GPT proof is an alternative martingale-coupling step, not the paper’s main Talagrand theorem.

AI involvementAI-centered
ImportanceModest
Technical verificationReported
Human understandingPartial

Weak Order on the MacNeille Completion of Bruhat Order

Colin Defant

  • Mathematics

Defant’s manuscript claims that, for every alternating sign matrix A, gradedness of [e,A], the corresponding equal-length condition and equidimensionality of all X_B in that interval are equivalent to Cohen–Macaulayness of all those ASM varieties. The new direction establishes Escobar–Klein–Weigandt Conjecture 3.21.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationReported
Human understandingFull

Weak Order on the MacNeille Completion of Bruhat Order

Colin Defant

  • Mathematics

Defant’s manuscript displays an interval in the weak order on six-by-six alternating sign matrices whose Möbius value is 1, although it has none of the specified parabolic forms. Its open-interval order complex is therefore not contractible, contrary to the Hamaker–Reiner dichotomy.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationReported
Human understandingFull

A Computably Enumerable tt-Degree Without Computably Enumerable Irreducible m-Degrees

Patrizio Cintioli

  • Mathematics
  • Theoretical Computer Science

Some computably enumerable truth-table degree contains no computably enumerable irreducible many-one degree. Cintioli derives the example using Degtev’s c.e. semirecursive set with rigid complement and the fact that its simple-set representatives cannot be irreducible cylinders.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Local Odd-Cycle Transversals in Generalized Mycielski Graphs and an Erdős Problem on Almost-Half Independent Sets

Przemek Chojecki, GPT-5.5 Pro

  • Mathematics
  • Theoretical Computer Science

An infinite-chromatic graph has the property that, for every nonnegative function f(m) tending to infinity, every m-vertex finite subgraph contains an independent set of size at least m/2−f(m). A stronger construction bounds local odd-cycle-transversal number by any prescribed unbounded integer profile.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

On the largest sum-free subset of the lattice cube

Peter Keevash, Jeck Lim

  • Mathematics

For every d≥1 and nonnegative r,s,t with r+s+t=dn, the uniform distributions on the lattice-cube slices L_d(r), L_d(s), and L_d(t) are jointly mixable. Keevash and Lim use this theorem to obtain the optimal limiting sum-free density of [n+1]^d and an O(n^(d−1)) upper error.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

Primitive sets and von Mangoldt chains: Erdős Problem #1196 and beyond

Boris Alexeev, Kevin Barreto, Yanyang Li, Jared Duker Lichtman, Liam Price, Jibran Iqbal Shah, Quanyu Tang, Terence Tao

  • Mathematics

For k≥1, a primitive set A of integers with at least k prime factors, and any set Q of odd primes, the sum f(A(Q)) is at most f(Nₖ(Q)), where f sums 1/(a log a).

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Primitive sets and von Mangoldt chains: Erdős Problem #1196 and beyond

Boris Alexeev, Kevin Barreto, Yanyang Li, Jared Duker Lichtman, Liam Price, Jibran Iqbal Shah, Quanyu Tang, Terence Tao

  • Mathematics

If A has positive upper doubly logarithmic density Δ, then A contains an infinite strictly increasing divisibility chain whose upper count through x, divided by log log x, is at least Δ.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

QED: An Open-Source Multi-Agent System for Generating Mathematical Proofs on Open Problems

Chenyang An, Qihao Ye, Minghao Pan, Jiayaun Zhang

  • Mathematics

For the discrete-time switch–walk–switch walk on Z₂ wr Z, started with all lamps off and walker positions 0 or 2, the total-variation distance between the time-t laws is comparable to t^(−1/2), with constants independent of t.

AI involvementAI-autonomous
ImportanceModest
Technical verificationFormal (human)
Human understandingFew

Global Product Intersection Sets in Semigroups

Wouter van Doorn, Pietro Monticone, Quanyu Tang

  • Mathematics
  • Theoretical Computer Science

For asymptotically strictly decreasing sequences, and for every index set Q with at least two members, the possible global product-intersection sets are exactly the subsets of positive integers containing 1. The paper also specifies the Q=0 and Q=1 cases.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

Max Cut with Small-Dimensional SDP Solutions

Hsien-Chih Chang, Suprovat Ghoshal, Euiwoong Lee

  • Mathematics
  • Theoretical Computer Science

For a feasible d-dimensional Max-Cut SDP solution strengthened by triangle inequalities, a randomized polynomial-time rounding algorithm attains expected cut value at least (α_GW+2^(−O(d))) times the SDP value.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Short proofs in combinatorics, probability and number theory II

Boris Alexeev, Moe Putterman, Mehtaab Sawhney, Mark Sellke, Gregory Valiant

  • Mathematics
  • Theoretical Computer Science

For every m≥1, an explicit K₄-free graph on 20m+31 vertices is 4-chromatic and vertex-critical, while every cycle has at most ten chords. This refutes the proposed unbounded forced chord count under the stated local 3-colorability condition.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Short proofs in combinatorics, probability and number theory II

Boris Alexeev, Moe Putterman, Mehtaab Sawhney, Mark Sellke, Gregory Valiant

  • Mathematics

There is no universal Erdős–Turán angular-discrepancy bound of order √(ν(f) log M(f)) with polynomial degree replaced by the number ν(f) of nonzero terms. For each N, a polynomial with ν=N+2 and M<3 has a positive real zero of multiplicity N+1, giving discrepancy at least N+1/2 on the stated short arc.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Short proofs in combinatorics, probability and number theory II

Boris Alexeev, Moe Putterman, Mehtaab Sawhney, Mark Sellke, Gregory Valiant

  • Mathematics

A single real sequence has uniformly bounded partial exponential sums A_k= sup_N |Σ_{n<N} exp(2πikx_n)|=O(√(k log(2k))) for every positive integer k. This implies the weaker limsup Ã_k=o(k) sought in Erdős Problem 987.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

An Improved Last-Iterate Convergence Rate for Anchored Gradient Descent Ascent

Anja Surina, Arun Suggala, George Tsoukalas, Anton Kovsharov, Sergey Shirobokov, Francisco J. R. Ruiz, Pushmeet Kohli, Swarat Chaudhuri

  • Mathematics
  • Theoretical Computer Science

For smooth convex-concave objectives with a monotone K-Lipschitz gradient operator and a saddle point, the schedule αₜ=1/(K√(t+γ)), βₜ=γ/(t+γ), γ≥2, gives squared-gradient residual O(1/t) at the last iterate of anchored gradient descent-ascent.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (Lean 4)
Human understandingFull

Short proofs in combinatorics and number theory

Boris Alexeev, Moe Putterman, Mehtaab Sawhney, Mark Sellke, Gregory Valiant

  • Mathematics

An explicit basis A of order two has the property that every partition A=A₁⊔A₂ leaves at least one of A₁+A₁ or A₂+A₂ with arbitrarily long gaps. The construction uses powers-of-five blocks and special points that force missing intervals.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationReported
Human understandingFull

The Simplicity of the Hodge Bundle

Anand Patel

  • Mathematics

Patel’s revised paper claims that for every genus g≥2, the Hodge bundle on the moduli stack of smooth complex genus-g curves has no nontrivial subbundle. The proof uses traces of involutions and curves with commuting involutions to constrain a hypothetical subbundle’s rank.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationPartial
Human understandingFull

Semi-Autonomous Formalization of the Vlasov-Maxwell-Landau Equilibrium

Vasily Ilin

  • Mathematics
  • Theoretical Physics

For smooth positive steady single-species Vlasov–Maxwell–Landau solutions on the three-torus with Coulomb collisions, Schwartz velocity decay and a polynomial score bound, the stated Lean theorem yields a spatially uniform Maxwellian at ion density, zero electric field and constant magnetic field.

AI involvementAI-led
ImportanceNotable
Technical verificationFormal (Lean 4 + human)
Human understandingFull

Transposition is Nearly Optimal for IID List Update

Christian Coester

  • Mathematics
  • Theoretical Computer Science

In the IID request model, the stationary expected access cost of Transposition is at most OPT+1, where OPT orders items by decreasing request probability. GPT-5 Pro proposed the central inequality sⱼ≤pⱼ and a coefficient-nonnegativity route; Christian Coester supplied the final proof.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

Early Evidence of Vibe-Proving with Consumer LLMs: A Case Study on Spectral Region Characterization with ChatGPT-5.2 (Thinking)

Brecht Verbeken, Brando Vagenende, Marie-Anne Guerry, Andres Algaba, Vincent Ginis

  • Mathematics
  • Theoretical Computer Science

For the specified row-stochastic 4×4 matrices supported on a 4-cycle, every nonreal eigenvalue a+ib has 0≤a≤1, a+|b|≤1 and (b²+a²+a)²+2a²−b²≥0. Conversely, upper-half-plane points with b>0, 0≤a≤1, a+b<1 and the polynomial nonnegative occur; the stated boundary arcs are attained. This includes Ran and Teng’s Conjecture 20.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

An Erdős problem on random subset sums in finite abelian groups

Jie Ma, Quanyu Tang

  • Mathematics

A repaired GPT-5.2 Pro argument uses prime cyclic groups to refute f(N)≤log₂N+o(log log N) for the random subset-sum covering threshold. Ma and Tang later prove the stronger lower bound f(p)≥log₂p+(1/(2 log 2)+o(1))log log p; the constant and quantitative refinement are their work.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Fel's Conjecture on Syzygies of Numerical Semigroups

Evan Chen, Chris Cummins, GSM, Dejan Grubisic, Leopold Haller, Letong Hong, Andranik Kurghinyan, Kenny Lau, Hugh Leather, Seewoo Lee, Aram Markosyan, Ken Ono, Manooshree Patel, Gaurang Pendharkar, Vedant Rathi, Alex Schneidman, Volker Seeker, Shubho Sengupta, Ishan Sinha, Jimmy Xin, Jujian Zhang

  • Mathematics
  • Theoretical Computer Science

For every numerical semigroup S and integer p≥0, AxiomProver’s Lean proof establishes Fel’s formula for the normalized alternating syzygy quantity Kₚ(S) in terms of universal symmetric polynomials, gap power sums and a final polynomial term.

AI involvementAI-autonomous
ImportanceNotable
Human understandingFull

Parity of k-differentials in genus zero and one

Dawei Chen, Evan Chen, Kenny Lau, Ken Ono, Jujian Zhang

  • Mathematics
  • Theoretical Computer Science

For odd k≥3 and integers n with gcd(n,k)=gcd(n+1,k)=1, AxiomProver found a Jacobi-symbol reformulation and the floor-sum identity N_k(n)=F_k(n+1)−F_k(n). The authors combine that identity with classical Jacobi facts to establish the parity formula N_k(n)≡⌊(k+1)/4⌋ mod 2 and make the earlier genus-zero and genus-one k-differential formulas unconditional.

AI involvementAI-autonomous
ImportanceNotable
Technical verificationPartial
Human understandingFull

Lower bounds for multivariate independence polynomials and their generalisations

Joonkyung Lee, Jaehyeon Seo

  • Mathematics

Lee and Seo prove a degreewise lower bound for the multivariate independence polynomial at nonnegative vertex weights and a two-proper-colour semiproper partition-function bound. The model-origin Lemma 2.1 and dual-set ideas are retained central steps; the equality characterization in their first theorem is outside the Lean target.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erdős Problems

Tony Feng, Trieu Trinh, Garrett Bingham, Jiwon Kang, Shengtong Zhang, Sang-hyun Kim, Kevin Barreto, Carl Schildkraut, Junehyuk Jung, Jaehyeon Seo, Carlo Pagano, Yuri Chervonyi, Dawsen Hwang, Kaiying Hou, Sergei Gukov, Cheng-Chiang Tsai, Hyunwoo Choi, Youngbeom Jin, Wei-Yuan Li, Hao-An Wu, Ruey-An Shiu, Yu-Sheng Shih, Quoc V. Le, Thang Luong

  • Mathematics
  • Theoretical Computer Science

For increasing positive integers aₙ with liminf aₙ^(1/2ⁿ)>1, Aletheia’s accepted proof shows that Σₙ 1/(aₙaₙ₊₁) is irrational. This is the original Erdős #1051 statement.

AI involvementAI-autonomous
ImportanceModest
Human understandingFull

Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erdős Problems

Tony Feng, Trieu Trinh, Garrett Bingham, Jiwon Kang, Shengtong Zhang, Sang-hyun Kim, Kevin Barreto, Carl Schildkraut, Junehyuk Jung, Jaehyeon Seo, Carlo Pagano, Yuri Chervonyi, Dawsen Hwang, Kaiying Hou, Sergei Gukov, Cheng-Chiang Tsai, Hyunwoo Choi, Youngbeom Jin, Wei-Yuan Li, Hao-An Wu, Ruey-An Shiu, Yu-Sheng Shih, Quoc V. Le, Thang Luong

  • Mathematics

Aletheia constructed two closed, infinite, countable compact subsets of the complex plane with the same zero transfinite diameter but different values of the polynomial sublevel-area quantity μ(F).

AI involvementAI-autonomous
ImportanceModest
Technical verificationFormal (human)
Human understandingFull

Semi-Autonomous Mathematics Discovery with Gemini: A Case Study on the Erdős Problems

Tony Feng, Trieu Trinh, Garrett Bingham, Jiwon Kang, Shengtong Zhang, Sang-hyun Kim, Kevin Barreto, Carl Schildkraut, Junehyuk Jung, Jaehyeon Seo, Carlo Pagano, Yuri Chervonyi, Dawsen Hwang, Kaiying Hou, Sergei Gukov, Cheng-Chiang Tsai, Hyunwoo Choi, Youngbeom Jin, Wei-Yuan Li, Hao-An Wu, Ruey-An Shiu, Yu-Sheng Shih, Quoc V. Le, Thang Luong

  • Mathematics

Aletheia constructed planar n-point sets with no four cocircular in which every point determines fewer than 3n/4 distinct distances. These refute the near-n-distances question under that hypothesis.

AI involvementAI-autonomous
ImportanceModest
Technical verificationFormal (human)
Human understandingFull

Solving a Research Problem in Mathematical Statistics with AI Assistance

Edgar Dobriban

  • Mathematics

For independent W(q,r) contamination and the stated Hölder-smooth clean and contaminated densities in p dimensions, revised Theorem 4.1 gives the squared-loss pointwise minimax rate max{n^(−2s/(2s+p)), ε^(2s/(s+1+p/r))}. A smoothed kernel estimator attains the upper bound.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (human)
Human understandingFull

Olympiad-level formal mathematical reasoning with reinforcement learning

Thomas Hubert, Rishi Mehta, Laurent Sartran, Miklós Z. Horváth, Goran Žužić, Eric Wieser, Aja Huang, Julian Schrittwieser, Yannick Schroecker, Hussain Masoom, Ottavia Bertolli, Tom Zahavy, Amol Mandhane, Jessica Yung, Iuliya Beloshapka, Borja Ibarz, Vivek Veeriah, Lei Yu, Oliver Nash, Paul Lezeau, Salvatore Mercuri, Calle Sönne, Bhavik Mehta, Alex Davies, Daniel Zheng, Fabian Pedregosa, Yin Li, Ingrid von Glehn, Mark Rowland, Samuel Albanie, Ameya Velingker, Simon Schmitt, Edward Lockhart, Edward Hughes, Henryk Michalewski, Nicolas Sonnerat, Demis Hassabis, Pushmeet Kohli, David Silver

  • Mathematics
  • Theoretical Computer Science

DeepMind reports Lean proofs for IMO 2024 problems 1, 2, and 6, and publishes the specific proof artifacts.

AI involvementAI-centered
ImportanceModest
Technical verificationFormal (Lean)
Human understandingFull

Point Convergence of Nesterov's Accelerated Gradient Method: An AI-Assisted Proof

Uijeong Jang, Ernest K. Ryu

  • Mathematics
  • Theoretical Computer Science

For a differentiable convex function with a minimizer, trajectories of the r=3 Nesterov ODE converge to a minimizer. Jang and Ryu adapt this argument to show that both iterates xₖ and yₖ of standard finite-dimensional Nesterov accelerated gradient converge to one minimizer for an L-smooth convex function under the stated growing tₖ schedule.

AI involvementAI-centered
ImportanceMajor
Technical verificationFormal (human)
Human understandingFull

Discovery of Unstable Singularities

Yongji Wang, Mehdi Bennani, James Martens, Sébastien Racanière, Sam Blackwell, Alex Matthews, Stanislav Nikolov, Gonzalo Cao-Labora, Daniel S. Park, Martin Arjovsky, Daniel Worrall, Chongli Qin, Ferran Alet, Borislav Kozlovskii, Nenad Tomašev, Alex Davies, Pushmeet Kohli, Tristan Buckmaster, Bogdan Georgiev, Javier Gómez-Serrano, Ray Jiang, Ching-Yao Lai

  • Mathematics
  • Theoretical Physics

Wang and collaborators report a second unstable self-similar Córdoba–Córdoba–Fontelos (CCF) profile, stable and three unstable incompressible porous media (IPM) profiles, and stable and three unstable 2D Boussinesq profiles. They test residuals and linearized modes on dense grids. A fourth Boussinesq profile remains an unvalidated candidate.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

The Oddtown problem modulo a composite number

Boris Bukh, Ting-Wei Chao, Zeyu Zheng

  • Mathematics

For every fixed modulus ℓ with at least two distinct prime factors, the AI-origin construction gives fℓ(n)≥ω(ℓ)n−o(n) through projective-hyperplane complements and low-rank correction blocks. This refutes fℓ(n)≤n outside prime powers.

AI involvementAI-centered
ImportanceNotable
Technical verificationReported
Human understandingFull

AlphaEvolve: A coding agent for scientific and algorithmic discovery

Alexander Novikov, Ngân Vũ, Marvin Eisenberger, Emilien Dupont, Po-Sen Huang, Adam Zsolt Wagner, Sergey Shirobokov, Borislav Kozlovskii, Francisco J. R. Ruiz, Abbas Mehrabian, M. Pawan Kumar, Abigail See, Swarat Chaudhuri, George Holland, Alex Davies, Sebastian Nowozin, Pushmeet Kohli, Matej Balog

  • Mathematics
  • Theoretical Computer Science

The AlphaEvolve report gives a rank-48 decomposition of the 4×4 matrix-multiplication tensor over the complex numbers, yielding a bilinear algorithm using 48 scalar multiplications.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFull

AlphaEvolve: A coding agent for scientific and algorithmic discovery

Alexander Novikov, Ngân Vũ, Marvin Eisenberger, Emilien Dupont, Po-Sen Huang, Adam Zsolt Wagner, Sergey Shirobokov, Borislav Kozlovskii, Francisco J. R. Ruiz, Abbas Mehrabian, M. Pawan Kumar, Abigail See, Swarat Chaudhuri, George Holland, Alex Davies, Sebastian Nowozin, Pushmeet Kohli, Matej Balog

  • Mathematics

AlphaEvolve’s authors report an eleven-dimensional kissing configuration with 593 outer spheres.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFew

AlphaGeometry solves 25 translated Olympiad geometry problems

  • Mathematics
  • Theoretical Computer Science

AlphaGeometry solved 25/30 specialized IMO-AG-30 classical Euclidean geometry translations using learned auxiliary-point constructions plus a symbolic deduction and algebra engine; the paper does not claim all natural-language IMO geometry conditions are represented.

AI involvementAI-centered
ImportanceNotable
Human understandingFull

Mathematical discoveries from program search with large language models

Bernardino Romera-Paredes, Mohammadamin Barekatain, Alexander Novikov, Matej Balog, M. Pawan Kumar, Emilien Dupont, Francisco J. R. Ruiz, Jordan S. Ellenberg, Pengming Wang, Omar Fawzi, Pushmeet Kohli, Alhussein Fawzi

  • Mathematics
  • Theoretical Computer Science

The authors report packing heuristics that outperform tested baselines on their online one-dimensional bin-packing experiments.

AI involvementAI-centered
ImportanceModest
Technical verificationPartial
Human understandingFull

Mathematical discoveries from program search with large language models

Bernardino Romera-Paredes, Mohammadamin Barekatain, Alexander Novikov, Matej Balog, M. Pawan Kumar, Emilien Dupont, Francisco J. R. Ruiz, Jordan S. Ellenberg, Pengming Wang, Omar Fawzi, Pushmeet Kohli, Alhussein Fawzi

  • Mathematics

The FunSearch paper reports new large cap-set constructions improving the authors’ stated finite-dimensional and asymptotic baselines.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFull

Faster sorting algorithms discovered using deep reinforcement learning

Daniel J. Mankowitz, Andrea Michi, Anton Zhernov, Marco Gelmi, Marco Selvi, Cosmin Paduraru, Edouard Leurent, Shariq Iqbal, Jean-Baptiste Lespiau, Alex Ahern, Thomas Köppe, Kevin Millikin, Stephen Gaffney, Sophie Elster, Jackson Broshear, Chris Gamble, Kieran Milan, Robert Tung, Minjae Hwang, Taylan Cemgil, Mohammadamin Barekatain, Yujia Li, Amol Mandhane, Thomas Hubert, Julian Schrittwieser, Demis Hassabis, Pushmeet Kohli, Martin Riedmiller, Oriol Vinyals, David Silver

  • Mathematics
  • Theoretical Computer Science

DeepMind reports AI-discovered assembly sorting routines translated to C++ and incorporated in LLVM libc++. Its tests report up to 70% improvement for length-five inputs and about 1.7% for inputs exceeding 250,000 elements.

AI involvementAI-centered
ImportanceNotable
Technical verificationPartial
Human understandingFull

Discovering faster matrix multiplication algorithms with reinforcement learning

Alhussein Fawzi, Matej Balog, Aja Huang, Thomas Hubert, Bernardino Romera-Paredes, Mohammadamin Barekatain, Alexander Novikov, Francisco J. R. Ruiz, Julian Schrittwieser, Grzegorz Swirszcz, David Silver, Demis Hassabis, Pushmeet Kohli

  • Mathematics
  • Theoretical Computer Science

The authors report an algorithm multiplying 4×4 matrices over the two-element field using 47 scalar multiplications, compared with 49 from two levels of Strassen’s algorithm.

AI involvementAI-centered
ImportanceNotable
Technical verificationFormal (exact computation)
Human understandingFull

Understanding the topics

Read the background

AI contributions across the mathematics fields in the catalog, including algebra, analysis, probability, combinatorics, geometry, topology, number theory, algorithms, optimization, and formal proof.

Following this subject includes its nested topics and problems. Contributions that belong to several fields are counted once within the subject.

Background review: 2026-10-14. Selected sources:

Website timeline

Publication dates and editorial changes.