OpenAlex ·
A Dimension-Free Bound on the Poincaré Constant of Isotropic Log-Concave Measures
- Published
- Source
- OpenAlex
- Published in
- arXiv (Cornell University)
- DOI
- 10.48550/arxiv.2610.07728
- Original
- Read it at the publisher
- Topics
Abstract
The Kannan--Lovász--Simonovits (KLS) conjecture asserts that isotropic log-concave probability measures have Poincaré constants bounded by a universal constant, independently of dimension. In this work, we provide a deterministic variational proof that the Poincaré constant $C_P(μ)\le25$, where $μ$ is any isotropic log-concave probability measure. The proof centers on mean-zero integration of compatible symmetric tensor fields. Starting from constant tensors, we obtain factorially normalized Appell polynomials by iterating this procedure. Using moment estimates and a quadratic variance inequality, we establish geometric bounds on these iterated primitives through a \emph{joint} variational argument over the measure and test function. At a maximizing pair, stationarity in the test function controls the weighted sum of the two highest-order cumulant terms arising from the measure variation. To extend these bounds to arbitrary compatible fields, we use a weighted Helmholtz--Hodge decomposition of weighted divergence, the formal adjoint of differentiation. Controlling the curl in the decomposition yields a curvature comparison for the integration operators, uniformly in tensor order. Combining this comparison with the Appell bounds gives a bound on the full integration operator and hence the stated Poincaré inequality. The estimate is independent of the auxiliary positive curvature, allowing approximation to general isotropic log-concave measures.
The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.
Read it on ReadSearch
Each opens after you create an account, and brings you back to this paper.
More in Numerical Analysis
A Neural JKO Scheme for Hellinger-Kantorovich Gradient Flows via Monge-Growth Pairs
Particle-Number-Preserving Quantum Signal Processing for Thermal Properties: Canonical-Ensemble Estimation and Grand-Canonical Reconstruction
Learned Adaptive Multiresolution Diffusion Imaging
An extended deep energy method for thermo-mechanical crack propagation
FOSLS-deRhaNN: native de Rham neural classes for H(div) and H(curl) with applications to first-order system least-squares neural network methods for partial differential equations
Get the few papers worth your time in Numerical Analysis every morning.
ReadSearch reads the new papers in your field each day and brings you the few worth reading, each with its key finding in a sentence.