A note on ubiquity of geometric Brascamp-Lieb data

Neal Bez,Anthony Gauvan,Hiroshi Tsuji

2024-07-31

Abstract:Relying substantially on work of Garg, Gurvits, Oliveira and Wigderson, it is shown that geometric Brascamp--Lieb data are, in a certain sense, ubiquitous. This addresses a question raised by Bennett and Tao in their recent work on the adjoint Brascamp--Lieb inequality.

Classical Analysis and ODEs

What problem does this paper attempt to address?

Adjoint Brascamp–Lieb inequalities

Jonathan Bennett,Terence Tao

DOI: https://doi.org/10.1112/plms.12633

2024-09-16

Proceedings of the London Mathematical Society

Abstract:Abstract The Brascamp–Lieb inequalities are a generalization of the Hölder, Loomis–Whitney, Young, and Finner inequalities that have found many applications in harmonic analysis and elsewhere. In this paper, we introduce an “adjoint” version of these inequalities, which can be viewed as an version of the entropic Brascamp–Lieb inequalities of Carlen and Cordero–Erausquin. As applications, we reprove a log‐convexity property of the Gowers uniformity norms, and establish some reverse inequalities for various tomographic transforms. We conclude with some open questions.

mathematics
Brascamp--Lieb inequalities on locally compact abelian groups

Jonathan Bennett,Michael G. Cowling

2024-12-16

Abstract:We establish a structure theorem for the Brascamp--Lieb constant formulated in the general setting of locally compact abelian groups. This extends and unifies the finiteness characterisations previously known for euclidean spaces and for finitely generated groups and their duals. We place particular emphasis on Fourier invariance throughout, reflecting the fundamental Fourier invariance of Brascamp--Lieb multilinear forms in this context.

Functional Analysis
Regularized Brascamp--Lieb inequalities

Neal Bez,Shohei Nakamura

2024-11-29

Abstract:Given any (forward) Brascamp--Lieb inequality on euclidean space, a famous theorem of Lieb guarantees that gaussian near-maximizers always exist. Recently, Barthe and Wolff used mass transportation techniques to establish a counterpart to Lieb's theorem for all non-degenerate cases of the inverse Brascamp--Lieb inequality. Here we build on work of Chen--Dafnis--Paouris and employ heat-flow techniques to understand the inverse Brascamp--Lieb inequality for certain regularized input functions, in particular extending the Barthe--Wolff theorem to such a setting. Inspiration arose from work of Bennett, Carbery, Christ and Tao for the forward inequality, and we recover their generalized Lieb's theorem using a clever limiting argument of Wolff. In fact, we use Wolff's idea to deduce regularized inequalites in the broader framework of the forward-reverse Brascamp--Lieb inequality, in particular allowing us to recover the gaussian saturation property in this framework first obtained by Courtade, Cuff, Liu and Verdú.

Classical Analysis and ODEs,Functional Analysis
Information-Theoretic Perspectives on Brascamp-Lieb Inequality and Its Reverse

Jingbo Liu,Thomas A. Courtade,Paul Cuff,Sergio Verdu

DOI: https://doi.org/10.48550/arXiv.1702.06260

2017-12-03

Abstract:We introduce an inequality which may be viewed as a generalization of both the Brascamp-Lieb inequality and its reverse (Barthe's inequality), and prove its information-theoretic (i.e.\ entropic) formulation. This result leads to a unified approach to functional inequalities such as the variational formula of Rényi entropy, hypercontractivity and its reverse, strong data processing inequalities, and transportation-cost inequalities, whose utility in the proofs of various coding theorems has gained growing popularity recently. We show that our information-theoretic setting is convenient for proving properties such as data processing, tensorization, convexity (Riesz-Thorin interpolation) and Gaussian optimality. In particular, we elaborate on a "doubling trick" used by Lieb and Geng-Nair to prove several results on Gaussian optimality. Several applications are discussed, including a generalization of the Brascamp-Lieb inequality involving Gaussian random transformations, the determination of Wyner's common information of vector Gaussian sources, and the achievable rate region of certain key generation problems in the case of vector Gaussian sources.

Information Theory
The Brascamp-Lieb inequality in Convex Geometry and in the Theory of Algorithms

Károly J. Böröczky

2024-12-15

Abstract:The Brascamp-Lieb inequality in harmonic analysis was proved by Brascamp and Lieb in the rank one case in 1976, and by Lieb in 1990. It says that in a certain inequality, the optimal constant can be determined by checking the inequality for centered Gaussian distributions. It was Keith M Ball's pioneering work around 1990 that led to various applications of the inequality in Convex Geometry, and even in Discrete Geometry, like Brazitikos' quantitative fractional version of the Helly Theorem. On the other hand, determining the optimal constant and possible Gaussian extremizers for the Brascamp-Lieb inequality can be formulated as a problem in terms of positive definite matrices, and this problem has intimate links to the Theory of Algorithms.

Metric Geometry
Quiver Brascamp-Lieb inequalities

Nicholas Hu

2024-02-22

Abstract:We formulate generalized Brascamp-Lieb inequalities for representations of bipartite quivers and establish necessary and sufficient conditions for such inequalities. Notably, we show contra Lieb that solely considering Gaussians is insufficient in the quiver setting.

Classical Analysis and ODEs
Brascamp-Lieb Inequality and Its Reverse: An Information Theoretic View

Jingbo Liu,Thomas A. Courtade,Paul Cuff,Sergio Verdu

DOI: https://doi.org/10.48550/arXiv.1605.02818

2016-05-10

Abstract:We generalize a result by Carlen and Cordero-Erausquin on the equivalence between the Brascamp-Lieb inequality and the subadditivity of relative entropy by allowing for random transformations (a broadcast channel). This leads to a unified perspective on several functional inequalities that have been gaining popularity in the context of proving impossibility results. We demonstrate that the information theoretic dual of the Brascamp-Lieb inequality is a convenient setting for proving properties such as data processing, tensorization, convexity and Gaussian optimality. Consequences of the latter include an extension of the Brascamp-Lieb inequality allowing for Gaussian random transformations, the determination of the multivariate Wyner common information for Gaussian sources, and a multivariate version of Nelson's hypercontractivity theorem. Finally we present an information theoretic characterization of a reverse Brascamp-Lieb inequality involving a random transformation (a multiple access channel).

Information Theory
Geometric G-functions and Atypicality

David Urbanik

2023-08-21

Abstract:We describe a general method for giving $p$-adic interpretations of $G$-functions arising from degenerating periods of smooth projective algebraic varieties. Using this, we are able to implement a strategy due to André for bounding heights of moduli points where period functions acquire unusual algebraic relations. This leads to new results on Galois lower bounds for special moduli, and new cases of the Zilber-Pink conjecture. In particular, we establish the first Galois-orbit lower bounds on CM moduli in non-Shimura settings. As a more technical contribution, we introduce a refinement of the Pila-Zannier strategy capable of handling Zilber-Pink-type atypical intersection problems in arbitrary dimension and for arbitrary smooth projective families.

Algebraic Geometry,Number Theory
G-invariant Hilbert schemes on Abelian surfaces and enumerative geometry of the orbifold Kummer surface

Stephen Pietromonaco

DOI: https://doi.org/10.1007/s40687-021-00298-9

2021-12-07

Research in the Mathematical Sciences

Abstract:For an Abelian surface A with a symplectic action by a finite group G, one can define the partition function for G-invariant Hilbert schemes ZA,G(q)=∑d=0∞e(Hilbd(A)G)qd.\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\begin{aligned} Z_{A, G}(q) = \sum _{d=0}^{\infty } e(\text {Hilb}^{d}(A)^{G})q^{d}. \end{aligned}$$\end{document}We prove the reciprocal ZA,G-1$$Z_{A,G}^{-1}$$ is a modular form of weight 12e(A/G)$$\frac{1}{2}e(A/G)$$ for the congruence subgroup Γ0(|G|)$$\Gamma _{0}(|G|)$$ and give explicit expressions in terms of eta products. Refined formulas for the χy$$\chi _{y}$$-genera of Hilb(A)G$$\text {Hilb}(A)^{G}$$ are also given. For the group generated by the standard involution τ:A→A$$\tau : A \rightarrow A$$, our formulas arise from the enumerative geometry of the orbifold Kummer surface [A/τ]$$[A/\tau ]$$. We prove that a virtual count of curves in the stack is governed by χy(Hilb(A)τ)$$\chi _{y}(\text {Hilb}(A)^{\tau })$$. Moreover, the coefficients of ZA,τ$$Z_{A, \tau }$$ are true (weighted) counts of rational curves, consistent with hyperelliptic counts of Bryan, Oberdieck, Pandharipande, and Yin.

mathematics
Information-theoretic inequalities on unimodular Lie groups

Gregory S Chirikjian,Gregory S. Chirikjian,

DOI: https://doi.org/10.3934/jgm.2010.2.119

2010-01-01

Journal of Geometric Mechanics

Abstract:Classical inequalities used in information theory such as those of de Bruijn, Fisher, Cramér, Rao, and Kullback carry over in a natural way from Euclidean space to unimodular Lie groups. These are groups that possess an integration measure that is simultaneously invariant under left and right shifts. All commutative groups are unimodular. And even in noncommutative cases unimodular Lie groups share many of the useful features of Euclidean space. The rotation and Euclidean motion groups, which are perhaps the most relevant Lie groups to problems in geometric mechanics, are unimodular, as are the unitary groups that play important roles in quantum computing. The extension of core information theoretic inequalities defined in the setting of Euclidean space to this broad class of Lie groups is potentially relevant to a number of problems relating to information gathering in mobile robotics, satellite attitude control, tomographic image reconstruction, biomolecular structure determination, and quantum information theory. In this paper, several definitions are extended from the Euclidean setting to that of Lie groups (including entropy and the Fisher information matrix), and inequalities analogous to those in classical information theory are derived and stated in the form of fifteen small theorems. In all such inequalities, addition of random variables is replaced with the group product, and the appropriate generalization of convolution of probability densities is employed. An example from the field of robotics demonstrates how several of these results can be applied to quantify the amount of information gained by pooling different sensory inputs.

mathematics, applied,physics, mathematical
Umbel convexity and the geometry of trees

F.P. Baudier,C. Gartland

DOI: https://doi.org/10.1016/j.aim.2023.109461

IF: 1.675

2024-02-01

Advances in Mathematics

Abstract:For every p ∈ ( 0 , ∞ ) , a new metric invariant called umbel p-convexity is introduced. The asymptotic notion of umbel convexity captures the geometry of countably branching trees, much in the same way as Markov convexity, the local invariant which inspired it, captures the geometry of bounded degree trees. Umbel convexity is used to provide a “Poincaré-type” metric characterization of the class of Banach spaces that admit an equivalent norm with Rolewicz's property (β). We explain how a relaxation of umbel p-convexity, called infrasup-umbel p-convexity, plays a role in obtaining compression rate bounds for coarse embeddings of countably branching trees. Local analogues of these invariants - fork p-convexity and infrasup-fork p-convexity - are introduced, and their relationship to Markov p-convexity and relaxations of the p-fork inequality is discussed. The metric invariants are estimated for a large class of Heisenberg groups, and in particular a parallelogram p-convexity inequality is proved for Heisenberg groups over p-uniformly convex Banach spaces. Finally, a new characterization of non-negative curvature is given.

mathematics
An Algebraic Brascamp-Lieb Inequality

Jennifer Duncan

DOI: https://doi.org/10.1007/s12220-021-00638-9

2024-01-16

Abstract:We prove a global nonlinear Brascamp-Lieb inequality for a general class of maps, encompassing polynomial and rational maps, as a consequence of the multilinear Kakeya-type inequalities of Zhang and Zorin-Kranich. We incorporate a natural affine-invariant weight that compensates for local degeneracies, and yields a uniform constant that depends only on the 'degree' of the maps involved.

Classical Analysis and ODEs
Cartan-Schouten metrics for information geometry and machine learning

Andre Diatta,Bakary Manga,Fatimata Sy

2024-08-28

Abstract:We study Cartan-Schouten metrics, explore invariant dual connections, and propose them as models for Information Geometry. Based on the underlying Riemannian barycenter and the biinvariant mean of Lie groups, we subsequently propose a new parametric mean for data science and machine learning which comes with several advantages compared to traditional tools such as the arithmetic mean, median, mode, expectation, least square method, maximum likelihood, linear regression. We call a metric on a Lie group, a Cartan-Schouten metric, if its Levi-Civita connection is biinvariant, so every 1-parameter subgroup through the unit is a geodesic. Except for not being left or right invariant in general, Cartan-Schouten metrics enjoy the same geometry as biinvariant metrics, since they share the same Levi-Civita connection. To bypass the non-invariance apparent drawback, we show that Cartan-Schouten metrics are completely determined by their value at the unit. We give an explicit formula for recovering them from their value at the unit, thus making them much less computationally demanding, compared to general metrics on manifolds. Furthermore, Lie groups with Cartan-Schouten metrics are complete Riemannian or pseudo-Riemannian manifolds. We give a complete characterization of Lie groups with Riemannian or Lorentzian Cartan-Schouten metrics. Cartan-Schouten metrics are in abundance on 2-nilpotent Lie groups. Namely, on every 2-nilpotent Lie group, there is a 1-1 correspondence between the set of left invariant metrics and that of Cartan-Schouten metrics.

Differential Geometry,Information Theory,Mathematical Physics,Probability,Statistics Theory
Unifying the Brascamp-Lieb Inequality and the Entropy Power Inequality

Venkat Anantharam,Varun Jog,Chandra Nair

DOI: https://doi.org/10.48550/arXiv.1901.06619

2021-09-27

Abstract:The entropy power inequality (EPI) and the Brascamp-Lieb inequality (BLI) are fundamental inequalities concerning the differential entropies of linear transformations of random vectors. The EPI provides lower bounds for the differential entropy of linear transformations of random vectors with independent components. The BLI, on the other hand, provides upper bounds on the differential entropy of a random vector in terms of the differential entropies of some of its linear transformations. In this paper, we define a family of entropy functionals, which we show are subadditive. We then establish that Gaussians are extremal for these functionals by mimicking the idea in Geng and Nair (2014). As a consequence, we obtain a new entropy inequality that generalizes both the BLI and EPI. By considering a variety of independence relations among the components of the random vectors appearing in these functionals, we also obtain families of inequalities that lie between the EPI and the BLI.

Information Theory,Probability
Geometric De Giorgi Theory

Lihe Wang

2023-08-08

Abstract:We prove De Giorgi-Nash-Moser Theory using a geometric approach.

Analysis of PDEs
New Brunn--Minkowski and functional inequalities via convexity of entropy

Gautam Aishwarya,Liran Rotem

2024-02-23

Abstract:We study the connection between the concavity properties of a measure $\nu$ and the convexity properties of the associated relative entropy $D(\cdot \Vert \nu)$ along optimal transport. As a corollary we prove a new dimensional Brunn-Minkowski inequality for centered star-shaped bodies, when the measure $\nu$ is log-concave with a p-homogeneous potential (such as the Gaussian measure). Our method allows us to go beyond the usual convexity assumption on the sets that is fundamentally essential for the standard differential-geometric technique in this area. We then take a finer look at the convexity properties of the Gaussian relative entropy, which yields new functional inequalities. First we obtain curvature and dimensional reinforcements to Otto--Villani's "HWI" inequality in the Gauss space, when restricted to even strongly log-concave measures. As corollaries, we obtain improved versions of Gross' logarithmic Sobolev inequality and Talgrand's transportation cost inequality in this setting.

Metric Geometry,Information Theory,Probability
GAIT: A Geometric Approach to Information Theory

Jose Gallego-Posada,Ankit Vani,Max Schwarzer,Simon Lacoste-Julien

DOI: https://doi.org/10.48550/arXiv.1906.08325

2020-10-14

Abstract:We advocate the use of a notion of entropy that reflects the relative abundances of the symbols in an alphabet, as well as the similarities between them. This concept was originally introduced in theoretical ecology to study the diversity of ecosystems. Based on this notion of entropy, we introduce geometry-aware counterparts for several concepts and theorems in information theory. Notably, our proposed divergence exhibits performance on par with state-of-the-art methods based on the Wasserstein distance, but enjoys a closed-form expression that can be computed efficiently. We demonstrate the versatility of our method via experiments on a broad range of domains: training generative models, computing image barycenters, approximating empirical measures and counting modes.

Machine Learning
Arithmeticity and geometrical commensurators

Yanlong Hao

2024-12-10

Abstract:This paper aims to characterize rank-one arithmetic and locally symmetric metrics in the coarsely geometric setting using coarse-geometric commensurators. We provide a positive answer in general under the Hilbert-Smith conjecture and unconditionally for finite volume negatively curved manifolds with finitely many cusps.

Geometric Topology
Pearcey universality at cusps of polygonal lozenge tiling

Jiaoyang Huang,Fan Yang,Lingfu Zhang

2023-08-15

Abstract:We study uniformly random lozenge tilings of general simply connected polygons. Under a technical assumption that is presumably generic with respect to polygon shapes, we show that the local statistics around a cusp point of the arctic curve converge to the Pearcey process. This verifies the widely predicted universality of edge statistics in the cusp case. Together with the smooth and tangent cases proved in Aggarwal-Huang and Aggarwal-Gorin, these are believed to be the three types of edge statistics that can arise in a generic polygon. Our proof is via a local coupling of the random tiling with non-intersecting Bernoulli random walks (NBRW). To leverage this coupling, we establish an optimal concentration estimate for the tiling height function around the cusp. As another step and also a result of potential independent interest, we show that the local statistics of NBRW around a cusp converge to the Pearcey process when the initial configuration consists of two parts with proper density growth, via careful asymptotic analysis of the determinantal formula.

Probability,Mathematical Physics,Combinatorics
Generalised Brègman relative entropies: a brief introduction

Ryszard Paweł Kostecki

2024-03-01

Abstract:We present some basic elements of the theory of generalised Brègman relative entropies over nonreflexive Banach spaces. Using nonlinear embeddings of Banach spaces together with the Euler--Legendre functions, this approach unifies two former approaches to Brègman relative entropy: one based on reflexive Banach spaces, another based on differential geometry. This construction allows to extend Brègman relative entropies, and related geometric and operator structures, to arbitrary-dimensional state spaces of probability, quantum, and postquantum theory. We give several examples, not considered previously in the literature.

Mathematical Physics,Information Theory

A note on ubiquity of geometric Brascamp-Lieb data

Adjoint Brascamp–Lieb inequalities

Brascamp--Lieb inequalities on locally compact abelian groups

Regularized Brascamp--Lieb inequalities

Information-Theoretic Perspectives on Brascamp-Lieb Inequality and Its Reverse

The Brascamp-Lieb inequality in Convex Geometry and in the Theory of Algorithms

Quiver Brascamp-Lieb inequalities

Brascamp-Lieb Inequality and Its Reverse: An Information Theoretic View

Geometric G-functions and Atypicality

G-invariant Hilbert schemes on Abelian surfaces and enumerative geometry of the orbifold Kummer surface

Information-theoretic inequalities on unimodular Lie groups

Umbel convexity and the geometry of trees

An Algebraic Brascamp-Lieb Inequality

Cartan-Schouten metrics for information geometry and machine learning

Unifying the Brascamp-Lieb Inequality and the Entropy Power Inequality

Geometric De Giorgi Theory

New Brunn--Minkowski and functional inequalities via convexity of entropy

GAIT: A Geometric Approach to Information Theory

Arithmeticity and geometrical commensurators

Pearcey universality at cusps of polygonal lozenge tiling

Generalised Brègman relative entropies: a brief introduction