Matching works

7 works

math.ST — Statistics Theory

The planar data-selection constants F(2,n)=2n/(2n−1)F(2,n)=2n/(2n-1), including F(2,3)=6/5F(2,3)=6/5 and F(2,2)=4/3F(2,2)=4/3

Contributed by Samuil Petkov

For a finite multiset D⊂R2D\subset\mathbb{R}^2 let LD∗L_D^* be the least mean squared distance from a point to DD, and let LD∗(n)L_D^*(n) be the least such loss attained at the average of nn points selected from DD with repetition. Hanneke, Moran, Shlimovich and Yehudayoff asked for F(d,n)=sup⁡DLD∗(n)/LD∗F(d,n)=\sup_D L_D^*(n)/L_D^*. We prove F(2,n)=2n/(2n−1)F(2,n)=2n/(2n-1) for every n≥2n\ge2: every finite planar multiset satisfies LD∗(n)≤2n2n−1LD∗L_D^*(n)\le\frac{2n}{2n-1}L_D^*, with equality for 2n−12n-1 copies of (0,0)(0,0) and one copy of (1,0)(1,0). In particular F(2,3)=6/5F(2,3)=6/5, and F(2,2)=4/3F(2,2)=4/3 receives a short proof. The proof reduces to three support points by a mean-preserving Carathéodory argument and then exhibits an explicit random selection of nn support points, supported on one small triangle of the selection grid, whose second moment about the population weights is dominated by 12n−1\frac{1}{2n-1} times the categorical covariance. For n∈{2,3}n\in\{2,3\} we also give random selections attaining this second moment exactly.

Author comments: 6 pages. Determines F(2,n) for all n, resolving the d=2 row of Question 1 of the COLT 2025 open problem "Data selection for regression tasks" (Hanneke, Moran, Shlimovich, Yehudayoff).

A mix of human-written and AI-generated textHuman understanding: all parts
Nov 16, 20266 pages
PDFView work

math.PR — Probability

Beyond Small Balls: A Determinant Inequality for Functions of Matrix Gaussian Series

Contributed by Suvrit Sra

We prove a functional extension of Gaussian operator-norm small-ball inequalities. Let X=∑i=1dgiBiX=\sum_{i=1}^d g_iB_i, where g∼N(0,Id)g\sim N(0,I_d), the matrices Bi∈Rd×dB_i\in\mathbb{R}^{d\times d} are symmetric, and ∑iBi2⪯Id\sum_iB_i^2\preceq I_d. For every even log-concave f ⁣:R→[0,∞)f\colon\mathbb{R}\to[0,\infty), we obtain Edet⁡f(X)≥(Ef(κG))d\mathbb{E}\det f(X)\ge(\mathbb{E} f(\kappa G))^d at dilation κ=13/5\kappa=13/5, where GG is standard Gaussian. The proof reduces each function to one small-ball estimate, obtained from matrix-weighted Poincaré inequalities, and controls the rest by averages over the spectrum. We give an equivalent entropy formulation. An explicit counterexample shows that no universal dilation below π/2\sqrt{\pi/2} can work; we conjecture that this value is optimal.

A mix of human-written and AI-generated textHuman understanding: all partsdeterminant inequalitylog-concave functionsmatrix Gaussian seriesrandom matricessmall-ball probabilities
Nov 02, 202630 pages
PDFView work

math.MG — Metric Geometry

Symmetric Volume Descent from U4 and a Four-Dimensional Lower Bound

Contributed by Hangang Wang

We explain a symmetric deformation that decreases the volume of the four-dimensional constant-width body U4U_4. The emphasis is on why the deformation works, rather than the size of the decrease. Oddness preserves width; convexity requires control of the curvature radii. A positive cap on one support branch is coupled to a negative correction on the adjacent branch. Their first-order volume contributions have different orders in the localization parameter, so the favorable contribution dominates. A small motion towards the ball makes the direction feasible for positive time. In a separate part we give a spectral lower bound for all four-dimensional bodies of constant width one, including an explicit positive refinement. A short projection argument using a three-dimensional theorem of Jiang gives the stronger lower bound 0.178722030.17872203. A substantial gap between the lower and upper bounds remains.

A mix of human-written and AI-generated textHuman understanding: all partsconstant widthspectral lower boundsvolume descent
Oct 29, 202616 pages
PDFView work

math.MG — Metric Geometry

Optimizing a cube–ball lower bound for the reverse Brunn–Minkowski constant

Contributed by Ricardo Pinho

We optimize a single term in the cube–Euclidean-ball construction of V. Patankar to obtain C_70 >= exp(1/(2e)) > 1.20194336847. The strongest public proposal located and reproduced in our dated source audit is Patankar's 1.201927, whose specified witness evaluates near 1.2019271425725. A separate finite-dimensional witness certifies 1.20194336. Two separately written arithmetic implementations, using Arb and rational logarithm/Stirling bounds respectively, verify the numerical comparisons and reject corrupted or excessive claims. The geometric reduction and passage to the limit are proved here. This unrefereed preprint is not peer reviewed and not formally verified; exhaustive priority is not asserted.

Primarily AI-generated textHuman understanding: some partscentrally symmetric convex bodiescube and Euclidean ballexact certificatesintrinsic volumesreverse Brunn–Minkowski inequalityrigorous arithmetic
Oct 26, 20263 pages
PDFView work

math.MG — Metric Geometry

An explicit finite cylinder cover of the regular tetrahedron saving 3.86×10−43.86\times10^{-4} below the half-area bound

Contributed by Ricardo Pinho

We give an explicit cover of a regular tetrahedron by 40964096 cylinders with compact triangular perpendicular bases. The sum of their base areas, divided by the tetrahedron's minimum orthogonal projection area, lies in [0.499613830783714148, 0.499613830783714149)[0.499613830783714148,\,0.499613830783714149). Thus the saving below 1/21/2 exceeds 0.0003861692162858510.000386169216285851. The construction uses a finite tilt in the angular-sector framework of OpenAI's September 2026 preprint and rationally specified radial cutoffs. Coverage is certified by exact linear-programming dual witnesses, and independently verified by exhaustive rational polyhedron clipping. We describe the geometric reduction, the certificate obligations, and reproducible verification of the full closed tetrahedron and the perpendicular base areas. No optimality is asserted.

Primarily AI-generated textHuman understanding: some parts
Oct 17, 20265 pages
PDFView work

math.MG — Metric Geometry

Breaking the Fourth Wall: The Planar Centroid Banach–Mazur Diameter Is Strictly Less than 44

Contributed by Scott Kominers

We prove that the centroid Banach–Mazur diameter of planar convex bodies is strictly less than 44, (slightly) improving the prior upper bound 69/17≈4.0588269/17\approx 4.05882 due to Lassak. The proof combines a sharp covariance-ellipse sandwich, whose equality cases in any single direction occur only for triangles, with a compactness argument: an extremal pair at distance 44 would consist of two triangles, which are linearly equivalent and hence must in fact be at distance 11. The resulting gap below 44 is uniform but nonquantitative.

A mix of human-written and AI-generated textHuman understanding: all partsBanach–Mazur diameter
Sep 28, 202611 pages
PDFView work

Advanced search

Text
Human understanding
Linked formalizations