Limit Distributions and Sensitivity Analysis for Entropic Optimal Transport on Countable Spaces

For probability measures supported on countable spaces we derive limit distributions for empirical entropic optimal transport quantities. In particular, we prove that the corresponding plan converges weakly to a centered Gaussian process. Furthermore, its optimal value is shown to be asymptotically normal. The results are valid for a large class of ground cost functions and generalize recently obtained limit laws for empirical entropic optimal transport quantities on finite spaces. Our proofs are based on a sensitivity analysis with respect to a weighted l-norm relying on the dual formulation of entropic optimal transport as well as necessary and sufficient optimality conditions for the entropic transport plan. This can be used to derive weak convergence of the empirical entropic optimal transport plan and value that results in weighted BorisovDudley-Durst conditions on the underlying probability measures. The weights are linked to an exponential penalty term for dual entropic optimal transport and the underlying ground cost function under consideration. Finally, statistical applications, such as bootstrap, are discussed.

[1]  R. Bass.  Convergence of probability measures , 2011 .

[2]  Marco Cuturi,et al.  Sinkhorn Distances: Lightspeed Computation of Optimal Transport , 2013, NIPS.

[3]  Jérémie Bigot,et al.  Central limit theorems for entropy-regularized optimal transport on finite spaces and statistical applications , 2017, Electronic Journal of Statistics.

[4]  Jason Altschuler,et al.  Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration , 2017, NIPS.

[5]  Shun-ichi Amari,et al.  Information Geometry for Regularized Optimal Transport and Barycenters of Patterns , 2019, Neural Computation.

[6]  Chong Chen,et al.  Spatiotemporal imaging with diffeomorphic optimal transportation , 2020, Inverse Problems.

[7]  D. Bertsekas,et al.  The auction algorithm for the transportation problem , 1989 .

[8]  Richard Sinkhorn.  A Relationship Between Arbitrary Positive Matrices and Doubly Stochastic Matrices , 1964 .

[9]  A. Antos,et al.  Convergence properties of functional estimates for discrete distributions , 2001 .

[10]  Lénaïc Chizat,et al.  Scaling Algorithms for Unbalanced Transport Problems , 2016, 1607.05816.

[11]  Dimitri P. Bertsekas,et al.  A new algorithm for the assignment problem , 1981, Math. Program..

[12]  Dirk A. Lorenz,et al.  Entropic regularization of continuous optimal transport problems , 2019, 1906.01333.

[13]  Amol Sasane.  A Friendly Approach to Functional Analysis , 2017 .

[14]  J. Yukich.  Metric entropy and the central limit theorem in Banach spaces , 1986 .

[15]  E. J. McShane,et al.  Extension of range of functions , 1934 .

[16]  Jonathan Weed,et al.  Statistical bounds for entropic optimal transport: sample complexity and the central limit theorem , 2019, NeurIPS.

[17]  Alexander Gasnikov,et al.  Computational Optimal Transport: Complexity by Accelerated Gradient Descent Is Better Than by Sinkhorn's Algorithm , 2018, ICML.

[18]  Nicolas Papadakis,et al.  Regularized Optimal Transport and the Rot Mover's Distance , 2016, J. Mach. Learn. Res..

[19]  Roberto Cominetti,et al.  Asymptotic analysis of the exponential penalty trajectory in linear programming , 1994, Math. Program..

[20]  Richard Sinkhorn.  Diagonal equivalence to matrices with prescribed row and column sums. II , 1967 .

[21]  Lénaïc Chizat,et al.  Supplementary Material Supplementary material for the paper : “ Faster Wasserstein Distance Estimation with the Sinkhorn Divergence , 2022 .

[22]  Max Sommerfeld,et al.  Inference for empirical Wasserstein distances on finite spaces , 2016, 1610.03287.

[23]  Marco Cuturi,et al.  Computational Optimal Transport: With Applications to Data Science , 2019 .

[24]  C. Villani.  Optimal Transport: Old and New , 2008 .

[25]  Arthur Cayley,et al.  The Collected Mathematical Papers: On Monge's “Mémoire sur la théorie des déblais et des remblais” , 2009 .

[26]  L. Kantorovich.  On the Translocation of Masses , 2006 .

[27]  Yin Tat Lee,et al.  Path Finding Methods for Linear Programming: Solving Linear Programs in Õ(vrank) Iterations and Faster Algorithms for Maximum Flow , 2014, 2014 IEEE 55th Annual Symposium on Foundations of Computer Science.

[28]  I. Csiszár.  $I$-Divergence Geometry of Probability Distributions and Minimization Problems , 1975 .

[29]  S. Rachev,et al.  Mass transportation problems , 1998 .

[30]  A. Galichon,et al.  Optimal Transport Methods in Economics , 2016 .

[31]  Victor M. Panaretos,et al.  Fréchet means and Procrustes analysis in Wasserstein space , 2017, Bernoulli.

[32]  Axel Munk,et al.  Optimal Transport: Fast Probabilistic Approximation with Exact Solvers , 2018, J. Mach. Learn. Res..

[33]  A. Munk,et al.  Limit Laws for Empirical Optimal Solutions in Stochastic Linear Programs , 2020, 2007.13473.

[34]  S. Evans,et al.  The phylogenetic Kantorovich–Rubinstein metric for environmental sequence samples , 2010, Journal of the Royal Statistical Society. Series B, Statistical methodology.

[35]  Gabriel Peyré,et al.  Sample Complexity of Sinkhorn Divergences , 2018, AISTATS.

[36]  Axel Munk,et al.  Empirical Regularized Optimal Transport: Statistical Theory and Applications , 2018, SIAM J. Math. Data Sci..

[37]  Y. Brenier.  Polar Factorization and Monotone Rearrangement of Vector-Valued Functions , 1991 .

[38]  F. Santambrogio.  Optimal Transport for Applied Mathematicians: Calculus of Variations, PDEs, and Modeling , 2015 .

[39]  James B. Orlin,et al.  A faster strongly polynomial minimum cost flow algorithm , 1993, STOC '88.

[40]  O. Smolyanov,et al.  The theory of differentiation in linear topological spaces , 1967 .

[41]  I. Borisov.  Problem of accuracy of approximation in the central limit theorem for empirical measures , 1983 .

[42]  Thomas M. Cover,et al.  Elements of Information Theory (Wiley Series in Telecommunications and Signal Processing) , 2006 .

[43]  Jon A. Wellner,et al.  Weak Convergence and Empirical Processes: With Applications to Statistics , 1996 .

[44]  Kei Kobayashi,et al.  Entropy-Regularized Optimal Transport on Multivariate Normal and q-normal Distributions , 2020, Entropy.

[45]  Werner Römisch,et al.  Delta Method, Infinite Dimensional , 2006 .

[46]  Max Sommerfeld,et al.  Empirical optimal transport on countable metric spaces: Distributional limits and statistical applications , 2017, The Annals of Applied Probability.

[47]  A. Munk,et al.  Colocalization for super-resolution microscopy via optimal transport , 2021, Nature Computational Science.

[48]  Alain Trouvé,et al.  Interpolating between Optimal Transport and MMD using Sinkhorn Divergences , 2018, AISTATS.

[49]  Jonathan Weed,et al.  An explicit analysis of the entropic penalty in linear programming , 2018, COLT.

[50]  P. Rigollet,et al.  Optimal-Transport Analysis of Single-Cell Gene Expression Identifies Developmental Trajectories in Reprogramming , 2019, Cell.