Matching works

5 works

math.PR — Probability

The Wasserstein cost of the Girsanov Theorem

Contributed by Simon Coste

This is an expository note on the asymptotics of the Wasserstein distance between the empirical measure of infinite-dimensional random paths and their common distribution. We present the classical proof for these asymptotics in the case of the Wiener measure, unifying quantization arguments from Dereich and coauthors or Boissard and Le Gouic, and we adapt it for It\^o diffusions sampled through the Girsanov theorem. More precisely, let Q\mathbb{Q} be the law of the solution of a stochastic differential equation with a bounded Lipschitz drift. Let W1,…,WnW_1,\ldots,W_n be independent Brownian paths, and let ZiZ_i be their Girsanov weights for Q\mathbb{Q}. The reweighted empirical measure $$\widehat\mu_n=\frac{\sum_{i=1}^n Z_i\delta_{W_i}}{\sum_{i=1}^n Z_i}$$ approximates Q\mathbb{Q} when nn is large. We prove that the Wasserstein cost Wpp(μ^n,Q)W_p^p(\widehat\mu_n,\mathbb{Q}) is of order (log⁡n)−p/2(\log n)^{-p/2}, in expectation and almost surely, for every fixed p⩾1p\geqslant1. In the Brownian case where Zi=1Z_i=1 these results show that log⁡nWp(μ^n,Q)\sqrt{\log n}W_p(\widehat\mu_n,\mathbb{Q}) is asymptotically sandwiched between 2λd\sqrt{2\lambda_d} and 42λd4\sqrt{2\lambda_d}, where λd\lambda_d is the first eigenvalue of the Laplacian operator in the unit ball. We provide numerical experiments to estimate the (conjectured) limiting constant in dimension d=1d=1.

Author comments: 16 pages. Full Lean formalization publicly available at gitlab.com/SimonCoste/wasserstein

Primarily human-written textHuman understanding: all parts
Nov 12, 202616 pages
PDFView work

math-ph — Mathematical Physics

Notes on Wasserstein distance and the JT gravitational ensemble

Contributed by Ville Samuli Keränen

This note completes the study of the Boltzmann–Wasserstein (BW) distance, an optimal-transport metric on quantum theories, begun in a companion paper (arXiv:2606.01296). There the distance was evaluated between macroscopically different theories; here we take the opposite limit, two independent draws of the JT-gravity ensemble, which are identical at the classical level. The distance is computable and is controlled by the connected spectral form factor: the ramp, the double-trumpet wormhole, regulated by the plateau, the spectral discreteness. The mean is exponentially small, ⟨W2⟩∼S0 e−S0\langle \mathcal{W}^2\rangle \sim S_0\, e^{-S_0}, whereas uncorrelated (Poisson) levels would give an unsuppressed O(1)O(1) distance: the self-distance is a sharp diagnostic of spectral chaos. As a metric, the BW distance renders the ensemble a point cloud: exponentially small, strongly anisotropic, its leading principal axis a coherent “breathing” deformation of the low-energy spectrum. Its effective dimension grows asymptotically as Deff≃0.42 S02D_{\mathrm{eff}} \simeq 0.42\, S_0^2, so the distance self-averages, its relative width shrinking as std/mean∼1/S0\mathrm{std}/\mathrm{mean} \sim 1/S_0. At low temperatures it localizes on the Airy edge, where a single Tracy–Widom mode dominates and self-averaging fails. We test the picture with exact-diagonalization GUE and SYK simulations: in GUE the effective dimension grows logarithmically at every temperature and is predicted at finite NN, with no free parameters, by the same covariance; unfolded SYK exhibits the predicted anisotropy and mode structure of the cloud. Together with the companion result, this fixes the resolution of the metric: the microstate rigidity computed here is the noise floor beneath a macroscopic difference between theories.

Primarily AI-generated textHuman understanding: some parts
Nov 11, 202637 pages
PDFView work

math.PR — Probability

The Ultimate Matching Theorem

Contributed by Michael Goldman, Francesco Grotto, Dario Trevisan

Let X1,…,XNX_1,\dots,X_N be independent uniform points in the unit square, matched either to an evenly spread deterministic configuration or to a second independent sample. For exponents α1,α2>2\alpha_1,\alpha_2>2 with 1/α1+1/α2=1/21/\alpha_1+1/\alpha_2=1/2, we prove that with probability tending to one a single permutation controls the jjth coordinate displacements in the exponential Orlicz norm of exponent αj\alpha_j, at the scale log⁡N/N\sqrt{\log N/N}, for both j=1,2j=1,2. This settles Talagrand's Ultimate Matching Conjecture for all pairs of finite critical exponents. The proof constructs, on the flat torus, a current whose divergence is Gaussian white noise projected on a dyadic grid and whose coordinates have the critical anisotropic exponential Orlicz bounds, and transfers it to matching through a primal optimal-transport reduction. At the hard endpoints (2,∞)(2,\infty) and (∞,2)(\infty,2) the conjecture fails: a Brownian local-time construction shows that the normalized optimal matching cost diverges in probability.

Primarily AI-generated textHuman understanding: some parts
Nov 11, 202670 pages
PDFView work

math.FA — Functional Analysis

The pp-Bures–Wasserstein distance via 3×33\times3 block matrices (with application to pp-Bures–Wasserstein barycenters)

Contributed by Suvrit Sra

We study the pp-Bures–Wasserstein distance Dp(A,B)=(tr Ap+tr Bp−2∥A1/2B1/2∥pp)1/2D_p(A,B)=(\mathrm{tr}\,A^p+\mathrm{tr}\,B^p-2\|A^{1/2}B^{1/2}\|_p^p)^{1/2} on positive semidefinite matrices; it equals the Bures–Wasserstein distance for p=1p=1, and the Hilbert–Schmidt distance for p=2p=2. We prove that DpD_p is a metric for every 1≤p≤21\le p\le2, by proving a stronger 2014 conjecture of Lin and van den Driessche on 3×33\times3 block matrices. Our main tool is a partition inequality for Schatten norms, proved by complex interpolation. We also characterize Fréchet means for DpD_p by a fixed-point equation and prove geometric convergence of the associated iteration.

A mix of human-written and AI-generated textHuman understanding: all partsBures-Wasserstein distanceLin-van den Driessche conjectureSchatten normbarycenterscomplex interpolationpartition inequalitypositive semidefinite block matrices
Nov 02, 202620 pages
PDFView work

math.PR — Probability

On the Asymptotic W1W_1 Cost of Two-Dimensional Semi-Discrete Matching

Contributed by Heng Ma

Let X1,X2,…X_1,X_2,\ldots be independent uniform points in the unit square and let λ\lambda denote Lebesgue probability measure. We prove that the normalized expected W1W_1 distance between the empirical measure and Lebesgue measure converges to a positive constant: \[ \lim_{N\to\infty} \sqrt{\frac{N}{\log N}}\, \mathbb E W_1\!\left(\frac1N\sum_{i=1}^N\delta_{X_i},\lambda\right) =c_\star. \] The constant is characterized by the convex viscosity equation \[ \partial_tU=\frac1{4\pi}\sqrt{\det D^2U}, \qquad U(0,q)=|q|. \] Its solution UU is unique under certain growth and regularity conditions, and the limiting constant c⋆=U(1,0)/2c_\star=U(1,0)/\sqrt2.

Primarily AI-generated textHuman understanding: no parts
Sep 30, 202646 pages
PDFView work

Advanced search

Text
Human understanding
Linked formalizations