Adaptive optimal transport

DOI10.1093/IMAIAI/IAZ008MaRDI QIDQ5006511zbMATH OpenOpenAlexFDO

Authors Montacer Essid, Esteban G. Tabak, Debra F. Laefer

Publication date 16 August 2021

Published in Information and Inference: A Journal of the IMA (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1807.00393

zbMATH Keywords

entropyminimaxoptimal transport

Mathematics Subject Classification ID

Optimal transportation (49Q22) Statistical aspects of information-theoretic topics (62B10)

Abstract: An adaptive, adversarial methodology is developed for the optimal transport problem between two distributions $mu$ and $u$, known only through a finite set of independent samples $(x_i{)}_{i=\mathrm{1..}N}$ and $(y_j{)}_{j=\mathrm{1..}M}$. The methodology automatically creates features that adapt to the data, thus avoiding reliance on a priori knowledge of data distribution. Specifically, instead of a discrete point-bypoint assignment, the new procedure seeks an optimal map $T(x)$ defined for all $x$, minimizing the Kullback-Leibler divergence between $(T(xi))$ and the target $(y_j)$. The relative entropy is given a sample-based, variational characterization, thereby creating an adversarial setting: as one player seeks to push forward one distribution to the other, the second player develops features that focus on those areas where the two distributions fail to match. The procedure solves local problems matching consecutive, intermediate distributions between $mu$ and $u$. As a result, maps of arbitrary complexity can be built by composing the simple maps used for each local problem. Displaced interpolation is used to guarantee global from local optimality. The procedure is illustrated through synthetic examples in one and two dimensions.

Recommendations

• Data-driven optimal transport
• Sample-based optimal transport and barycenter problems
• Computational optimal transport. With applications to data sciences
• Preconditioning of optimal transport
• Optimal transport for generative models

Cited in

(15)

• Multi-projection of unequal dimension optimal transport theory for generative adversary networks
• Supervised learning of sheared distributions using linearized optimal transport
• A particle-evolving method for approximating the optimal transport plan
• An implicit gradient-descent procedure for minimax problems
• Data-driven optimal transport
• Explanation of variability and removal of confounding factors from data through optimal transport
• A distributed framework for the construction of transport maps
• Sample-based optimal transport and barycenter problems
• Preconditioning of optimal transport
• Learning to match via inverse optimal transport
• Linear optimal transport embedding: provable Wasserstein classification for certain rigid transformations and perturbations
• <scp>SISTA</scp>: Learning Optimal Transport Costs under Sparsity Constraints
• Computational methods for adapted optimal transport
• Mirror variational transport: a particle-based algorithm for distributional optimization on constrained domains
• A geometric view of optimal transportation and generative model

This page was built for publication: Adaptive optimal transport

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5006511)