Strong equivalence between metrics of Wasserstein type

Erhan Bayraktar,Gaoyue Guo

2024-06-29

Abstract:We appended an errata to the original submission. The purpose of this errata is to point out two errors in [2] and give a weakened version of those statements made.

Probability,Functional Analysis

What problem does this paper attempt to address?

Correction to "Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations"

Daniel Paulin,Peter A. Whalley

2024-02-15

Abstract:A method for analyzing non-asymptotic guarantees of numerical discretizations of ergodic SDEs in Wasserstein-2 distance is presented by Sanz-Serna and Zygalakis in ``Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations". They analyze the UBU integrator which is strong order two and only requires one gradient evaluation per step, resulting in desirable non-asymptotic guarantees, in particular $\mathcal{O}(d^{1/4}\epsilon^{-1/2})$ steps to reach a distance of $\epsilon > 0$ in Wasserstein-2 distance away from the target distribution. However, there is a mistake in the local error estimates in Sanz-Serna and Zygalakis (2021), in particular, a stronger assumption is needed to achieve these complexity estimates. This note reconciles the theory with the dimension dependence observed in practice in many applications of interest.

Machine Learning,Numerical Analysis,Probability
Correction to: Two weight Sobolev norm inequalities for smooth Calderón–Zygmund operators and doubling weights

Eric T. Sawyer,Brett D. Wick

DOI: https://doi.org/10.1007/s00209-024-03515-7

2024-05-22

Mathematische Zeitschrift

Abstract:We correct two errors in the paper "Two weight Sobolev norm inequalities for smooth Calderón–Zygmund operators and doubling weights" published in Math. Z. Here we give a precise description of where the errors occur in the original paper, followed by an overview of the corrections, pointing to details appearing on the arXiv. We include statements of the main corrected theorems.

mathematics
Bounding adapted Wasserstein metrics

Jose Blanchet,Martin Larsson,Jonghwa Park,Johannes Wiesel

2024-07-31

Abstract:The Wasserstein distance $\mathcal{W}_p$ is an important instance of an optimal transport cost. Its numerous mathematical properties as well as applications to various fields such as mathematical finance and statistics have been well studied in recent years. The adapted Wasserstein distance $\mathcal{A}\mathcal{W}_p$ extends this theory to laws of discrete time stochastic processes in their natural filtrations, making it particularly well suited for analyzing time-dependent stochastic optimization problems. While the topological differences between $\mathcal{A}\mathcal{W}_p$ and $\mathcal{W}_p$ are well understood, their differences as metrics remain largely unexplored beyond the trivial bound $\mathcal{W}_p\lesssim \mathcal{A}\mathcal{W}_p$. This paper closes this gap by providing upper bounds of $\mathcal{A}\mathcal{W}_p$ in terms of $\mathcal{W}_p$ through investigation of the smooth adapted Wasserstein distance. Our upper bounds are explicit and are given by a sum of $\mathcal{W}_p$, Eder's modulus of continuity and a term characterizing the tail behavior of measures. As a consequence, upper bounds on $\mathcal{W}_p$ automatically hold for $\mathcal{AW}_p$ under mild regularity assumptions on the measures considered. A particular instance of our findings is the inequality $\mathcal{A}\mathcal{W}_1\le C\sqrt{\mathcal{W}_1}$ on the set of measures that have Lipschitz kernels. Our work also reveals how smoothing of measures affects the adapted weak topology. In fact, we find that the topology induced by the smooth adapted Wasserstein distance exhibits a non-trivial interpolation property, which we characterize explicitly: it lies in between the adapted weak topology and the weak topology, and the inclusion is governed by the decay of the smoothing parameter.

Probability,Statistics Theory
On the equivalence between Fourier-based and Wasserstein distances for probability measures on $\mathbb N$

Fei Cao,Xiaoqian Gong

2024-04-06

Abstract:In this manuscript we investigate the equivalence of Fourier-based metrics on discrete state spaces with the well-known Wasserstein distances. While the use of Fourier-based metrics in continuous state spaces is ubiquitous since its introduction by Giuseppe Toscani and his colleagues [9, 14, 16] in the study of kinetic-type partial differential equations, the introduction of its discrete analog is recent [2] and seems to be far less studied. In this work, various relations between Fourier-based metrics and Wasserstein distances are shown to hold when the state space is the set of non-negative integers $\mathbb N$. Lastly, we also describe potential applications of such equivalence of metrics in models from econophysics which motivate the present work.

Probability
Metric properties of partial and robust Gromov-Wasserstein distances

Jannatul Chhoa,Michael Ivanitskiy,Fushuai Jiang,Shiying Li,Daniel McBride,Tom Needham,Kaiying O'Hare

2024-11-04

Abstract:The Gromov-Wasserstein (GW) distances define a family of metrics, based on ideas from optimal transport, which enable comparisons between probability measures defined on distinct metric spaces. They are particularly useful in areas such as network analysis and geometry processing, as computation of a GW distance involves solving for registration between the objects which minimizes geometric distortion. Although GW distances have proven useful for various applications in the recent machine learning literature, it has been observed that they are inherently sensitive to outlier noise and cannot accommodate partial matching. This has been addressed by various constructions building on the GW framework; in this article, we focus specifically on a natural relaxation of the GW optimization problem, introduced by Chapel et al., which is aimed at addressing exactly these shortcomings. Our goal is to understand the theoretical properties of this relaxed optimization problem, from the viewpoint of metric geometry. While the relaxed problem fails to induce a metric, we derive precise characterizations of how it fails the axioms of non-degeneracy and triangle inequality. These observations lead us to define a novel family of distances, whose construction is inspired by the Prokhorov and Ky Fan distances, as well as by the recent work of Raghvendra et al.\ on robust versions of classical Wasserstein distance. We show that our new distances define true metrics, that they induce the same topology as the GW distances, and that they enjoy additional robustness to perturbations. These results provide a mathematically rigorous basis for using our robust partial GW distances in applications where outliers and partial matching are concerns.

Metric Geometry,Machine Learning
On properties of the Generalized Wasserstein distance

Benedetto Piccoli,Francesco Rossi

DOI: https://doi.org/10.48550/arXiv.1304.7014

2014-11-18

Abstract:The Wasserstein distances $W_p$ ($p\geq 1$), defined in terms of solution to the Monge-Kantorovich problem, are known to be a useful tool to investigate transport equations. In particular, the Benamou-Brenier formula characterizes the square of the Wasserstein distance $W_2$ as the infimum of the kinetic energy, or action functional, of all vector fields moving one measure to the other. Another important property of the Wasserstein distances is the Kantorovich-Rubinstein duality stating the equality between the distance $W_1$ and the supremum of the integrals of Lipschitz continuous functions with Lipschitz constant bounded by one. An intrinsic limitation of Wasserstein distances is the fact that they are defined only between measures having the same mass. To overcome such limitation, we recently introduced the generalized Wasserstein distances $W_p^{a,b}$, defined in terms of both the classical Wasserstein distance $W_p$ and the total variation (or $L^1$) distance. Here $p$ plays the same role as for the classic Wasserstein distance, while $a$ and $b$ are weights for the transport and the total variation term. In this paper we prove two important properties of the generalized Wasserstein distances: 1) a generalized Benamou-Brenier formula providing the equality between $W_2^{a,b}$ and the supremum of an action functional, which includes a transport term (kinetic energy) and a source term. 2) a duality à la Kantorovich-Rubinstein establishing the equality between $W_1^{1,1}$ and the flat metric.

Analysis of PDEs,Optimization and Control
Corrigendum to “Differentiating the stochastic entropy in negatively curved spaces under conformal changes”

François Ledrappier,Lin Shu

DOI: https://doi.org/10.5802/AIF.3389

2021-01-01

Abstract:We correct an error in [5, Proposition 2.16], and explain the subsequent changes in the proof of the main results.Résumé.-Nous corrigeons une erreur dans [5, Proposition 2.16] et nous détaillons les modifications à apporter aux démonstrations des principaux résultats.
On the metric property of quantum Wasserstein divergences

Gergely Bunth,József Pitrik,Tamás Titkos,Dániel Virosztek

DOI: https://doi.org/10.1103/PhysRevA.110.022211

2024-08-13

Abstract:Quantum Wasserstein divergences are modified versions of quantum Wasserstein distances defined by channels, and they are conjectured to be genuine metrics on quantum state spaces by De Palma and Trevisan. We prove triangle inequality for quantum Wasserstein divergences for every quantum system described by a separable Hilbert space and any quadratic cost operator under the assumption that a particular state involved is pure, and all the states have finite energy. We also provide strong numerical evidence suggesting that the triangle inequality holds in general, for an arbitrary choice of states.

Mathematical Physics,Functional Analysis,Quantum Physics
On nonexpansiveness of metric projection operators on Wasserstein spaces

Anshul Adve,Alpár Mészáros

2024-02-20

Abstract:In this paper we investigate properties of metric projections onto specific closed and geodesically convex proper subsets of Wasserstein spaces $(\mathcal{P}_p(\mathbf{R}^d),W_p).$ When $d=1$, as $(\mathcal{P}_2(\mathbf{R}),W_2)$ is isometrically isomorphic to a flat space with a Hilbertian structure, the corresponding projection operators are expected to be nonexpansive. We give a direct proof of this fact, relying on intrinsic analysis, which also implies nonexpansiveness in certain special cases in higher dimensions. When $d>1$, we show the failure of this property in two regimes: when $p>1$ is either small enough or large enough. Finally, we prove some positive curvature properties of Wasserstein spaces $(\mathcal{P}_p(\mathbf{R}^d),W_p)$ when $d\ge 2$ and $p\in(1,+\infty)$ are arbitrary: we show that Wasserstein spaces are nowhere locally Busemann NPC spaces, and they nowhere locally satisfy the so-called projection criterion. As a corollary of the former, they have nonnegative upper Alexandrov curvature, in a precise sense that we define here. In our analysis a particular subset of probability measures having densities uniformly bounded above by a given constant plays a special role.

Functional Analysis
Comparison of viscosity solutions for a class of second order PDEs on the Wasserstein space

Erhan Bayraktar,Ibrahim Ekren,Xin Zhang

2024-10-15

Abstract:We prove a comparison result for viscosity solutions of second order parabolic partial differential equations in the Wasserstein space. The comparison is valid for semisolutions that are Lipschitz continuous in the measure in a Fourier-Wasserstein metric and uniformly continuous in time. The class of equations we consider is motivated by Mckean-Vlasov control problems with common noise and filtering problems. The proof of comparison relies on a novel version of Ishii's lemma, which is tailor-made for the class of equations we consider.

Analysis of PDEs,Optimization and Control,Probability
On Vojta's $1+ε$ Conjecture

Xi Chen

DOI: https://doi.org/10.48550/arXiv.0705.1727

2007-05-11

Algebraic Geometry

Abstract:I gave a geometric proof of Vojta's 1 + epsilon conjecture. Some gaps in the published paper were spotted and kindly pointed out to me by Paul Vojta. These were addressed in "Erratum".
On metrics robust to noise and deformations

William Leeb

2024-05-06

Abstract:We study the properties of a family of distances between functions of a single variable. These distances are examples of integral probability metrics, and have been used previously for comparing probability measures on the line; special cases include the Earth Mover's Distance and the Kolmogorov Metric. We examine their properties for general signals, proving that they are robust to a broad class of deformations. We also establish corresponding robustness results for the induced sliced distances between multivariate functions. Finally, we establish error bounds for approximating the univariate metrics from finite samples, and prove that these approximations are robust to additive Gaussian noise. The results are illustrated in numerical experiments, which include comparisons with Wasserstein distances.

Functional Analysis,Numerical Analysis
Estimating processes in adapted Wasserstein distance

Julio Backhoff,Daniel Bartl,Mathias Beiglböck,Johannes Wiesel

DOI: https://doi.org/10.48550/arXiv.2002.07261

2021-05-17

Abstract:A number of researchers have independently introduced topologies on the set of laws of stochastic processes that extend the usual weak topology. Depending on the respective scientific background this was motivated by applications and connections to various areas (e.g. Plug-Pichler - stochastic programming, Hellwig - game theory, Aldous - stability of optimal stopping, Hoover-Keisler - model theory). Remarkably, all these seemingly independent approaches define the same adapted weak topology in finite discrete time. Our first main result is to construct an adapted variant of the empirical measure that consistently estimates the laws of stochastic processes in full generality. A natural compatible metric for the weak adapted topology is the given by an adapted refinement of the Wasserstein distance, as established in the seminal works of Pflug-Pichler. Specifically, the adapted Wasserstein distance allows to control the error in stochastic optimization problems, pricing and hedging problems, optimal stopping problems, etc. in a Lipschitz fashion. The second main result of this article yields quantitative bounds for the convergence of the adapted empirical measure with respect to adapted Wasserstein distance. Surprisingly, we obtain virtually the same optimal rates and concentration results that are known for the classical empirical measure wrt. Wasserstein distance.

Probability,Statistics Theory
Corrections to “Equivocations, Exponents and Second-Order Coding Rates under Various Rényi Information Measures”

Masahito Hayashi,Vincent Y. F. Tan

DOI: https://doi.org/10.1109/tit.2024.3355506

IF: 2.5

2024-01-01

IEEE Transactions on Information Theory

Abstract:There exists a gap for the proofs of the converse parts of Eqs. (49) and (50) in Theorem 1 and Eq. (74) of Theorem 3. These converse parts use Lemma 4, whose proofs contain a gap. We fix these errors when the rate is larger than the critical rate. A special case of Eq. (50) does not coincide the recent result (K. Li, and Y. Yao, arXiv:2209.00554v1) when the rate is smaller than the critical rate.

computer science, information systems,engineering, electrical & electronic
Note on the metric entropy for multivalued maps

Jan Andres,Pavel Ludvík

2024-06-24

Abstract:The main aim of this note is to point out by means of counter-examples that some arguments of the proofs of two theorems about a "half variational principle" for multivalued maps, formulated recently by Vivas and Sirvent [Metric entropy for set-valued maps, Discrete Contin. Dyn. Syst. Ser. B, 27 (2022), pp. 6589-6604], are false and that our corrected versions require rather restrictive additional assumptions. Nevertheless, we will be able to establish the full variational principle for a special subclass of multivalued lower semicontinuous maps with convex compact values on a compact subset of a Banach space.

Dynamical Systems
The Subfield Metric and its Application to Quantum Error Correction

Markus Grassl,Anna-Lena Horlemann,Violetta Weger

DOI: https://doi.org/10.1142/S021949882550063X

2024-05-24

Abstract:We introduce a new weight and corresponding metric over finite extension fields for asymmetric error correction. The weight distinguishes between elements from the base field and the ones outside of it, which is motivated by asymmetric quantum codes. We set up the theoretic framework for this weight and metric, including upper and lower bounds, asymptotic behavior of random codes, and we show the existence of an optimal family of codes achieving the Singleton-type upper bound.

Information Theory,Quantum Physics
Sliced Wasserstein Geodesics and Equivalence Wasserstein and Sliced Wasserstein metrics

John Seale Hopper

2024-07-10

Abstract:This paper will introduce a family of sliced Wasserstein geodesics which are not standard Wasserstein geodesics, objects yet to be discovered in the literature. These objects exhibit how the geometric structure of the Sliced Wasserstein space differs from the Wasserstein space, and provides a simple example of how solving the barycenter and gradient flow problems change when moving between these metrics. Some of these geodesics will only be Hölder continuous with respect to the Wasserstein metric and thus will provide a direct proof that Sliced-Wasserstein and regular Wasserstein metrics are not equivalent. Previous proofs of this were done for various cases in [2] and [5]. This paper, not only provides a direct proof, but also fills in gaps showing these metrics not equivalent in dimensions greater than 2.

Probability,Optimization and Control
Some notes on monotone set-valued measures and Egoroff's theorem

Jun Li

DOI: https://doi.org/10.1016/j.fss.2021.09.012

IF: 4.462

2022-02-01

Fuzzy Sets and Systems

Abstract:In this note, we point out that Theorem 3 (a version of Egoroff's theorem for monotone set-valued measures) shown in the paper Wu and Wu (2019) [25] is not valid, and we present two revised versions. We also point out that the proofs of Lemma 2 and Theorem 1 are defective, thus the truth of their conclusions cannot be affirmed. Thereby, an open problem for the characteristics of monotone measures is raised.

mathematics, applied,statistics & probability,computer science, theory & methods
Errata of the Unique Solvability of the Absolute Value Equation

Shi-Liang Wu,Peng Guo

DOI: https://doi.org/10.1007/s10957-024-02376-8

2024-01-28

Journal of Optimization Theory and Applications

Abstract:A published counterexample motivates us to correct one of our theorems in the previously published work in Wu and Guo (J Optim Theory Appl 169:705–712, 2016).

mathematics, applied,operations research & management science
Metric extrapolation in the Wasserstein space

Thomas O. Gallouët,Andrea Natale,Gabriele Todeschi

2024-07-22

Abstract:In this article we study a variational problem providing a way to extend for all times minimizing geodesics connecting two given probability measures, in the Wasserstein space. This is simply obtained by allowing for negative coefficients in the classical variational characterization of Wasserstein barycenters. We show that this problem admits two equivalent convex formulations: the first can be seen as a particular instance of Toland duality and the second is a barycentric optimal transport problem. We propose an efficient numerical scheme to solve this latter formulation based on entropic regularization and a variant of Sinkhorn algorithm.

Optimization and Control

Strong equivalence between metrics of Wasserstein type

Correction to "Wasserstein distance estimates for the distributions of numerical approximations to ergodic stochastic differential equations"

Correction to: Two weight Sobolev norm inequalities for smooth Calderón–Zygmund operators and doubling weights

Bounding adapted Wasserstein metrics

On the equivalence between Fourier-based and Wasserstein distances for probability measures on $\mathbb N$

Metric properties of partial and robust Gromov-Wasserstein distances

On properties of the Generalized Wasserstein distance

Corrigendum to “Differentiating the stochastic entropy in negatively curved spaces under conformal changes”

On the metric property of quantum Wasserstein divergences

On nonexpansiveness of metric projection operators on Wasserstein spaces

Comparison of viscosity solutions for a class of second order PDEs on the Wasserstein space

On Vojta's $1+ε$ Conjecture

On metrics robust to noise and deformations

Estimating processes in adapted Wasserstein distance

Corrections to “Equivocations, Exponents and Second-Order Coding Rates under Various Rényi Information Measures”

Note on the metric entropy for multivalued maps

The Subfield Metric and its Application to Quantum Error Correction

Sliced Wasserstein Geodesics and Equivalence Wasserstein and Sliced Wasserstein metrics

Some notes on monotone set-valued measures and Egoroff's theorem

Errata of the Unique Solvability of the Absolute Value Equation

Metric extrapolation in the Wasserstein space