Magnitude function determines generic finite metric spaces

Jun O'Hara

2024-05-11

Abstract:We give sufficient conditions for a finite metric space to be determined by the magnitude function. In particular, a generic finite metric space such that the distances between the points are rationally independent is determined by the asymptotic behavior of the magnitude function.

Metric Geometry

What problem does this paper attempt to address?

On the magnitude function of domains in Euclidean space

Heiko Gimperlein,Magnus Goffeng

DOI: https://doi.org/10.1353/ajm.2021.0023

2020-03-17

Abstract:We study Leinster's notion of magnitude for a compact metric space. For a smooth, compact domain $X\subset \mathbb{R}^{2m-1}$, we find geometric significance in the function $\mathcal{M}_X(R) = \mathrm{mag}(R\cdot X)$. The function $\mathcal{M}_X$ extends from the positive half-line to a meromorphic function in the complex plane. Its poles are generalized scattering resonances. In the semiclassical limit $R \to \infty$, $\mathcal{M}_X$ admits an asymptotic expansion. The three leading terms of $\mathcal{M}_X$ at $R=+\infty$ are proportional to the volume, surface area and integral of the mean curvature. In particular, for convex $X$ the leading terms are proportional to the intrinsic volumes, and we obtain an asymptotic variant of the convex magnitude conjecture by Leinster and Willerton, with corrected coefficients.

Metric Geometry,Analysis of PDEs,Classical Analysis and ODEs,Spectral Theory
Magnitude homology is a derived functor

Yasuhiko Asao,Sergei O. Ivanov

2024-05-20

Abstract:We prove that the magnitude (co)homology of an enriched category can, under some technical assumptions, be described in terms of derived functors between certain abelian categories. We show how this statement is specified for the cases of quasimetric spaces, finite quasimetric spaces, and finite digraphs. For quasimetric spaces, we define the notion of a distance module over a quasimetric space, define the functor of (co)invariants of a distance module and show that the magnitude (co)homology can be presented via its derived functors. As a corollary we obtain that the magnitude cohomology of a quasimetric space can be presented in terms of Ext functors in the category of distance modules. For finite quasimetric spaces, we show that magnitude (co)homology can be presented in terms of Tor and Ext functors over a certain graded algebra, called the distance algebra of the quasimetric space. For finite digraphs, the distance algebra is a bound quiver algebra. In addition, we show that the magnitude cohomology algebra of a finite quasimetric space can be described as a Yoneda algebra.

K-Theory and Homology
Metric Spaces of Arbitrary Finitely-Generated Scaling Group

Daniel N. Levitin

2023-06-20

Abstract:For a metric space $X$ with a compatible measure $\mu$, Genevois and Tessera defined the Scaling Group of $(X,\mu)$ as the subgroup $\Gamma$ of $\mathbb{R}_{>0}$ of positive real numbers $\gamma$ for which there are quasi-isometries of $X$ coarsely scaling $\mu$ by a factor of $\gamma$. We show that for any finitely generated subgroup $\Gamma$ of $\mathbb{R}_{>0}$ there exists a space $N_\Gamma$, bi-Lipschitz equivalent to a graph of finite degree, with scaling group $\Gamma$.

Metric Geometry,Group Theory
Magnitude and magnitude homology of filtered set enriched categories

Yasuhiko Asao

2023-07-25

Abstract:In this article, we give a framework for studying the Euler characteristic and its categorification of objects across several areas of geometry, topology and combinatorics. That is, the magnitude theory of filtered sets enriched categories. It is a unification of the Euler characteristic of finite categories and it the magnitude of metric spaces, both of which are introduced by Leinster. Our definitions cover a class of metric spaces which is broader than the original ones, so that magnitude (co)weighting of infinite metric spaces can be considered. We give examples of the magnitude from various research areas containing the Poincaré polynomial of ranked posets and the growth function of finitely generated groups. In particular, the magnitude homology gives categorifications of them. We also discuss homotopy invariance of the magnitude homology and its variants. Such a homotopy includes digraph homotopy and r-closeness of Lipschitz maps. As a benefit of our categorical view point, we generalize the notion of Grothendieck fibrations of small categories to our enriched categories, whose restriction to metric spaces is a notion called metric fibration that is initially introduced by Leinster. It is remarkable that the magnitude of such a fibration is a product of those of the fiber and the base. We especially study fibrations of graphs, and give examples of graphs with the same magnitude but are not isomorphic.

Algebraic Topology,Category Theory,Metric Geometry
The magnitude and spectral geometry

Heiko Gimperlein,Magnus Goffeng,Nikoletta Louca

2024-06-14

Abstract:We study the geometric significance of Leinster's notion of magnitude for a smooth manifold with boundary of arbitrary dimension, motivated by open questions for the unit disk in $\mathbb{R}^2$. For a large class of distance functions, including embedded submanifolds of Euclidean space and Riemannian manifolds satisfying a technical condition, we show that the magnitude function is well defined for $R\gg 0$ and admits a meromorphic continuation to sectors in $\mathbb{C}$. In the semiclassical limit $R \to \infty$, the magnitude function admits an asymptotic expansion, which determines the volume, surface area and integrals of generalized curvatures. Lower-order terms are computed by black box computer algebra. We initiate the study of magnitude analogues to classical questions in spectral geometry and prove an asymptotic variant of the Leinster-Willerton conjecture.

Differential Geometry,Analysis of PDEs,Classical Analysis and ODEs,Metric Geometry,Spectral Theory
Approximating Metric Magnitude of Point Sets

Rayna Andreeva,James Ward,Primoz Skraba,Jie Gao,Rik Sarkar

2024-09-07

Abstract:Metric magnitude is a measure of the "size" of point clouds with many desirable geometric properties. It has been adapted to various mathematical contexts and recent work suggests that it can enhance machine learning and optimization algorithms. But its usability is limited due to the computational cost when the dataset is large or when the computation must be carried out repeatedly (e.g. in model training). In this paper, we study the magnitude computation problem, and show efficient ways of approximating it. We show that it can be cast as a convex optimization problem, but not as a submodular optimization. The paper describes two new algorithms - an iterative approximation algorithm that converges fast and is accurate, and a subset selection method that makes the computation even faster. It has been previously proposed that magnitude of model sequences generated during stochastic gradient descent is correlated to generalization gap. Extension of this result using our more scalable algorithms shows that longer sequences in fact bear higher correlations. We also describe new applications of magnitude in machine learning - as an effective regularizer for neural network training, and as a novel clustering criterion.

Machine Learning,Metric Geometry
Asymptotic dimension and hyperfiniteness of generic Cantor actions

Sumun Iyer,Forte Shinko

2024-09-05

Abstract:We show that for a countable discrete group which is locally of finite asymptotic dimension, the generic continuous action on Cantor space has hyperfinite orbit equivalence relation. In particular, this holds for free groups, answering a question of Frisch-Kechris-Shinko-Vidnyánszky.

Logic,Dynamical Systems
Magnitude homology equivalence of Euclidean sets

Adrián Doña Mateo,Tom Leinster

2024-06-18

Abstract:Magnitude homology is an $\mathbf{R}^+$-graded homology theory of metric spaces that captures information on the complexity of geodesics. Here we address the question: when are two metric spaces magnitude homology equivalent, in the sense that there exist back-and-forth maps inducing mutually inverse maps in homology? We give a concrete geometric necessary and sufficient condition in the case of closed Euclidean sets. Along the way, we introduce the convex-geometric concepts of inner boundary and core, and prove a strengthening for closed convex sets of the classical theorem of Carathéodory.

Metric Geometry,Algebraic Topology,Category Theory
On metric spaces with given transfinite asymptotic dimensions

Yan Wu,Jingming Zhu,Taras Radul

DOI: https://doi.org/10.1016/j.topol.2022.108135

IF: 0.583

2022-06-01

Topology and its Applications

Abstract:For every countable ordinal number ξ, we construct a metric space Xξ whose transfinite asymptotic dimension and complementary-finite asymptotic dimension are both ξ.

mathematics, applied
Magnitudes, Scalable Monoids and Quantity Spaces

Dan Jonsson

DOI: https://doi.org/10.48550/arXiv.2108.02106

2023-01-11

Abstract:In ancient Greek mathematics, magnitudes such as lengths were strictly distinguished from numbers. In modern quantity calculus, a distinction is made between quantities and scalars that serve as measures of quantities. It can be argued that quantities should play a more prominent, independent role in modern mathematics, as magnitudes earlier. The introduction includes a sketch of the development and structure of the pre-modern theory of magnitudes and numbers. Then, a scalable monoid over a ring is defined and its basic properties are described. Congruence relations on scalable monoids, direct and tensor products of scalable monoids, subalgebras and homomorphic images of scalable monoids, and unit elements of scalable monoids are also defined and analyzed. A quantity space is defined as a commutative scalable monoid over a field, admitting a finite basis similar to a basis for a free abelian group. The mathematical theory of quantity spaces forms the basis of a rigorous quantity calculus and is developed with a view to applications in metrology and foundations of physics.

Rings and Algebras
Monotone Sequences of Metric Spaces with Compact Limits

R. Perales,C. Sormani

2024-06-27

Abstract:In this paper, we consider a fixed metric space (possibly an oriented Riemannian manifold with boundary) with an increasing sequence of distance functions and a uniform upper bound on diameter. When the metric space endowed with the pointwise limit of these distances is compact, then there is uniform and Gromov-Hausdorff (GH) convergence to this limit. When the metric space also has an integral current structure of uniformly bounded total mass (as is true for an oriented Riemannian manifold with boundary that has a uniform bound on total volume), we prove volume preserving intrinsic flat convergence to a subset of the GH limit whose closure is the whole GH limit. We provide a review of all notions and have a list of open questions at the end. Dedicated to Xiaochun Rong.

Metric Geometry,Differential Geometry
Magnitude homology and homotopy type of metric fibrations

Yasuhiko Asao,Yu Tajima,Masahiko Yoshinaga

2024-09-05

Abstract:In this article, we show that each two metric fibrations with a common base and a common fiber have isomorphic magnitude homology, and even more, the same magnitude homotopy type. That can be considered as a generalization of a fact proved by T. Leinster that the magnitude of a metric fibration with finitely many points is a product of those of the base and the fiber. We also show that the definition of the magnitude homotopy type due to the second and the third authors is equivalent to the geometric realization of Hepworth and Willerton's pointed simplicial set.

Algebraic Topology
Asymptotics of integral points, equivariant compactifications and equidistributions for homogeneous spaces

Runlin Zhang

2024-08-05

Abstract:Let U be a homogeneous variety over rational numbers of some linear algebraic group. Fix an integral model and assume the existence of infinitely many integral points. Then one would like to give an asymptotic count with the help of some height function. In many cases, with the help of measure rigidity of unipotent flows, we reduce this problem to one on equivariant birational geometry. For instance, we show that if G and H are both connected, semisimple, simply connected and without compact factors, then U=G/H is strongly Hardy-Littlewood with respect to certain height function. We also show that when H is "large" in H and both G and H are connected, reductive and without non-trivial Q-characters, then for every equivariant height, the asymptotic of integral points is the same as the volume asymptotic up to a constant. Three concrete examples with explicit heights are also provided to illustrate our approach.

Dynamical Systems,Algebraic Geometry,Number Theory
On isometric universality of spaces of metrics

Yoshito Ishiki,Katsuhisa Koshino

2024-09-26

Abstract:A metric space $(M, d)$ is said to be universal for a class of metric spaces if all metric spaces in the class can be isometrically embedded into $(M, d)$. In this paper, for a metrizable space $Z$ possessing abundant subspaces, we first prove that the space of bounded metrics on $Z$ is universal for all bounded metric spaces (with restricted cardinality). Next, in contrast, we show that if $Z$ is an infinite discrete space, then the space of metrics on $Z$ is universal for all separable metric spaces. As a corollary of our results, if $Z$ is non-compact, or uncountable and compact, then the space of metrics on $Z$ is universal for all compact metric spaces. In addition, if $Z$ is compact and countable, then there exists a compact metric space that can not be isometrically embedded into the space of metrics on $Z$.

Metric Geometry,Functional Analysis,General Topology
Finitude of physical measures for random maps

Pablo G. Barrientos,Fumihiko Nakamura,Yushi Nakano,Hisayoshi Toyokawa

2024-12-04

Abstract:For random compositions of independent and identically distributed measurable maps on a Polish space, we study the existence and finitude of absolutely continuous ergodic stationary probability measures (which are, in particular, physical measures) whose basins of attraction cover the whole space almost everywhere. We characterize and hierarchize such random maps in terms of their associated Markov operators, as well as show the difference between classes in the hierarchy by plenty of examples, including additive noise, multiplicative noise, and iterated function systems. We also provide sufficient practical conditions for a random map to belong to these classes. For instance, we establish that any continuous random map on a compact Riemannian manifold with absolutely continuous transition probability has finitely many physical measures whose basins of attraction cover Lebesgue almost all the manifold.

Dynamical Systems,Probability,Spectral Theory
An interpolation of metrics and spaces of metrics

Yoshito Ishiki

2024-10-18

Abstract:As a generalization of Hausdorff's extension theorem of metrics, we prove an interpolation theorem of a family of metrics defined on closed subsets of metrizable spaces. As an application, we investigate typicality of subsets of moduli spaces of metrics. We observe that various sets of all metrics with properties appearing in metric geometry are dense intersections of countable open subsets in spaces of metrics on metrizable spaces. For instance, our study is applicable to the set of all non-doubling metrics and the set of all non-uniformly disconnected metrics.

Metric Geometry,General Topology
Minimal projective resolution and magnitude homology of geodetic metric spaces

Yasuhiko Asao,Shun Wakatsuki

2024-08-22

Abstract:Asao-Ivanov showed that magnitude homology is a Tor functor, hence we can compute it by giving a projective resolution of a certain module. In this article, we compute magnitude homology by constructing a minimal projective resolution. As a consequence, we determine magnitude homology of geodetic metric spaces. We show that it is a free $\mathbb Z$-module, and give a recursive algorithm for constructing all cycles. As a corollary, we show that a finite geodetic metric space is diagonal if and only if it contains no 4-cuts. Moreover, we give explicit computations for cycle graphs, Petersen graph, Hoffman-Singleton graph, and a missing Moore graph. It includes another approach to the computation for cycle graphs, which has been studied by Hepworth--Willerton and Gu.

Algebraic Topology
Generic separable metric structures

Alexander Usvyatsov

DOI: https://doi.org/10.48550/arXiv.0802.0156

2008-02-02

Abstract:We compare three notions of genericity of separable metric structures. Our analysis provides a general model theoretic technique of showing that structures are generic in descriptive set theoretic (topological) sense and in measure theoretic sense. In particular, it gives a new perspective on Vershik's theorems on genericity and randomness of Urysohn's space among separable metric spaces.

Logic,General Topology
Majorizing Measures, Codes, and Information

Yifeng Chu,Maxim Raginsky

DOI: https://doi.org/10.48550/arXiv.2305.02960

2023-05-06

Abstract:The majorizing measure theorem of Fernique and Talagrand is a fundamental result in the theory of random processes. It relates the boundedness of random processes indexed by elements of a metric space to complexity measures arising from certain multiscale combinatorial structures, such as packing and covering trees. This paper builds on the ideas first outlined in a little-noticed preprint of Andreas Maurer to present an information-theoretic perspective on the majorizing measure theorem, according to which the boundedness of random processes is phrased in terms of the existence of efficient variable-length codes for the elements of the indexing metric space.

Information Theory,Probability,Machine Learning
Metric Space Magnitude for Evaluating the Diversity of Latent Representations

Katharina Limbeck,Rayna Andreeva,Rik Sarkar,Bastian Rieck

2024-10-31

Abstract:The magnitude of a metric space is a novel invariant that provides a measure of the 'effective size' of a space across multiple scales, while also capturing numerous geometrical properties, such as curvature, density, or entropy. We develop a family of magnitude-based measures of the intrinsic diversity of latent representations, formalising a novel notion of dissimilarity between magnitude functions of finite metric spaces. Our measures are provably stable under perturbations of the data, can be efficiently calculated, and enable a rigorous multi-scale characterisation and comparison of latent representations. We show their utility and superior performance across different domains and tasks, including (i) the automated estimation of diversity, (ii) the detection of mode collapse, and (iii) the evaluation of generative models for text, image, and graph data.

Machine Learning,Geometric Topology

Magnitude function determines generic finite metric spaces

On the magnitude function of domains in Euclidean space

Magnitude homology is a derived functor

Metric Spaces of Arbitrary Finitely-Generated Scaling Group

Magnitude and magnitude homology of filtered set enriched categories

The magnitude and spectral geometry

Approximating Metric Magnitude of Point Sets

Asymptotic dimension and hyperfiniteness of generic Cantor actions

Magnitude homology equivalence of Euclidean sets

On metric spaces with given transfinite asymptotic dimensions

Magnitudes, Scalable Monoids and Quantity Spaces

Monotone Sequences of Metric Spaces with Compact Limits

Magnitude homology and homotopy type of metric fibrations

Asymptotics of integral points, equivariant compactifications and equidistributions for homogeneous spaces

On isometric universality of spaces of metrics

Finitude of physical measures for random maps

An interpolation of metrics and spaces of metrics

Minimal projective resolution and magnitude homology of geodetic metric spaces

Generic separable metric structures

Majorizing Measures, Codes, and Information

Metric Space Magnitude for Evaluating the Diversity of Latent Representations