Abstract
The entropy-regularized Wasserstein transportation problem is useful for solving lots of problems in various applications. We study the deformed escort entropy regularization including the q-escort regularization and prove that the family of the inverse escort distributions of the optimal transportation plans forms a deformed exponential family, which has dually flat information-geometric structure. This elucidates the role of the escort transformation and its inverse in the theory of deformed exponential families. We further prove that the regularized cost function gives the dual potential of the flat manifold. We derive a new divergence function between two probability distributions in a probability simplex and a related Riemannian metric based on the regularized cost.
Similar content being viewed by others
References
Amari, S.: Information Geometry and Its Applications. Springer, Berlin (2016)
Amari, S., Karakida, R., Oizumi, M.: Information geometry connecting Wasserstein distance and Kullback–Leibler divergence via the entropy-relaxed transportation problem. Inf. Geom. 1, 13–37 (2018)
Amari, S., Karakida, R., Oizumi, M., Cuturi, M.: Information geometry for regularized optimal transport and barycenters of patterns. Neural Comput. 31, 827–848 (2019)
Amari, S., Ohara, A., Matsuzoe, H.: Geometry of deformed exponential families: invariant, dually-flat and conformal geometries. Phys. A 391, 4308–4319 (2012)
Ay, N., Jost, J., Le, H.V., Schwachhöfer, L.: Information Geometry. Springer, Berlin (2017)
Cuturi, M.: Sinkhorn distances: light speed computation of optimal transport. In: Advances in Neural Information Processing Systems 2292–2300 (2013)
Cuturi, M., Peyré, G.: A smoothed dual formulation for variational Wasserstein problems. SIAM J. Imaging Sci. 9 (2016)
Feydy, J., Sehourne, T., Vialard, F.-X., Amari, S., Trouve, A., Peyre, G.: Interpolating between optimal transport and MMD using Sinkhorn divergences. PMLR 89 (AISTATS’19), 2681–2690 (2019)
Frogner, C., Zhang, C., Mobahi, H., Araya-Polo, M., Poggio, T.: Learning with a Wasserstein loss. Adv. Neural Inf. Process. Syst. (NIPS) 28 (2015)
Genevay, A., Peyre, G., Cuturi, M.: Learning generative models with Sinkhorn divergences. PMLR 84 (AISTATS’18), 1608–1617 (2018)
Khan, G. and Zhang, J.: The Kähler geometry of certain optimal transport problems. ar**v:1812.00032v4 (2019)
Li, W., Zhao, Li.: Wasserstein information matrix (2020) (To appear)
Montavon, G., Muller, K., Cuturi, M.: Wasserstein training for Boltzmann machines. ar**v:1507.01972v1 (2015)
Muzellec, B., Nock, R., Patrini, G., Nielsen, F.: Tsallis regularized optimal transport and ecological inference. ar**v:1609.04495v1 (2016)
Naudts, J.: Estimators escort probabilities and \(phi\)-exponential families in statistical physics. J. Inequal. Pure Appl. Math. 5, 122 (2004)
Naudts, J.: Generalized Thermostatistics. Springer, Berlin (2011)
Naudts, J., Zhang, J.: \(\rho \)-\(\tau \) embedding and gauge freedom in information geometry. Inf. Geom. 1, 79–115 (2018)
Peyré, G., Cuturi, M.: Computational optimal transport: with applications to data science. Found. Trends Theor. Comput. Sci. 11, 355–607 (2019)
Santambrogio, F.: Optimal Transport for Applied Mathematicians. Birkhauser, Basel (2015)
Tsallis, C.: Introduction to Non-extensive Statistical Mechanics: Approaching a Complex World. Springer, Berlin (2009)
Villani, C.: Topics in Optimal Transportation. Graduate Studies in Mathematics. AMS, Providence (2003)
Villani, C.: Optimal Transport: Old and New. Springer, Berlin (2008)
Wong, T.-K.L.: Logarithmic divergence from optimal transport and Renyi geometry. Inf. Geom. 1, 39–78 (2018)
Wong, T.-K.L., Yang, J.: Pseudo-Riemannian geometry embeds information geometry in optimal transport. ar**v:1906.00030v4 (2021)
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Rights and permissions
About this article
Cite this article
Kurose, T., Yoshizawa, S. & Amari, Si. Optimal transportation plans with escort entropy regularization. Info. Geo. 5, 79–95 (2022). https://doi.org/10.1007/s41884-021-00058-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s41884-021-00058-2