Proceedings of the 36th International Conference on Machine Learning, PMLR 97:3982-3991, 2019.
Abstract
We provide theoretical analyses for two algorithms that solve the regularized optimal transport (OT) problem between two discrete probability measures with at most $n$ atoms. We show that a greedy variant of the classical Sinkhorn algorithm, known as the Greenkhorn algorithm, can be improved to $\bigOtil\left(n^2/\varepsilon^2\right)$, improving on the best known complexity bound of $\bigOtil\left(n^2/\varepsilon^3\right)$. This matches the best known complexity bound for the Sinkhorn algorithm and helps explain why the Greenkhorn algorithm outperforms the Sinkhorn algorithm in practice. Our proof technique is based on a primal-dual formulation and provide a tight upper bound for the dual solution, leading to a class of adaptive primal-dual accelerated mirror descent (APDAMD) algorithms. We prove that the complexity of these algorithms is $\bigOtil\left(n^2\sqrt{\gamma}/\varepsilon\right)$ in which $\gamma \in (0, n]$ refers to some constants in the Bregman divergence. Experimental results on synthetic and real datasets demonstrate the favorable performance of the Greenkhorn and APDAMD algorithms in practice.
@InProceedings{pmlr-v97-lin19a,
title = {On Efficient Optimal Transport: An Analysis of Greedy and Accelerated Mirror Descent Algorithms},
author = {Lin, Tianyi and Ho, Nhat and Jordan, Michael},
booktitle = {Proceedings of the 36th International Conference on Machine Learning},
pages = {3982--3991},
year = {2019},
editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan},
volume = {97},
series = {Proceedings of Machine Learning Research},
address = {Long Beach, California, USA},
month = {09--15 Jun},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v97/lin19a/lin19a.pdf},
url = {http://proceedings.mlr.press/v97/lin19a.html},
abstract = {We provide theoretical analyses for two algorithms that solve the regularized optimal transport (OT) problem between two discrete probability measures with at most $n$ atoms. We show that a greedy variant of the classical Sinkhorn algorithm, known as the \emph{Greenkhorn algorithm}, can be improved to $\bigOtil\left(n^2/\varepsilon^2\right)$, improving on the best known complexity bound of $\bigOtil\left(n^2/\varepsilon^3\right)$. This matches the best known complexity bound for the Sinkhorn algorithm and helps explain why the Greenkhorn algorithm outperforms the Sinkhorn algorithm in practice. Our proof technique is based on a primal-dual formulation and provide a \emph{tight} upper bound for the dual solution, leading to a class of \emph{adaptive primal-dual accelerated mirror descent} (APDAMD) algorithms. We prove that the complexity of these algorithms is $\bigOtil\left(n^2\sqrt{\gamma}/\varepsilon\right)$ in which $\gamma \in (0, n]$ refers to some constants in the Bregman divergence. Experimental results on synthetic and real datasets demonstrate the favorable performance of the Greenkhorn and APDAMD algorithms in practice.}
}
%0 Conference Paper
%T On Efficient Optimal Transport: An Analysis of Greedy and Accelerated Mirror Descent Algorithms
%A Tianyi Lin
%A Nhat Ho
%A Michael Jordan
%B Proceedings of the 36th International Conference on Machine Learning
%C Proceedings of Machine Learning Research
%D 2019
%E Kamalika Chaudhuri
%E Ruslan Salakhutdinov
%F pmlr-v97-lin19a
%I PMLR
%J Proceedings of Machine Learning Research
%P 3982--3991
%U http://proceedings.mlr.press
%V 97
%W PMLR
%X We provide theoretical analyses for two algorithms that solve the regularized optimal transport (OT) problem between two discrete probability measures with at most $n$ atoms. We show that a greedy variant of the classical Sinkhorn algorithm, known as the Greenkhorn algorithm, can be improved to $\bigOtil\left(n^2/\varepsilon^2\right)$, improving on the best known complexity bound of $\bigOtil\left(n^2/\varepsilon^3\right)$. This matches the best known complexity bound for the Sinkhorn algorithm and helps explain why the Greenkhorn algorithm outperforms the Sinkhorn algorithm in practice. Our proof technique is based on a primal-dual formulation and provide a tight upper bound for the dual solution, leading to a class of adaptive primal-dual accelerated mirror descent (APDAMD) algorithms. We prove that the complexity of these algorithms is $\bigOtil\left(n^2\sqrt{\gamma}/\varepsilon\right)$ in which $\gamma \in (0, n]$ refers to some constants in the Bregman divergence. Experimental results on synthetic and real datasets demonstrate the favorable performance of the Greenkhorn and APDAMD algorithms in practice.
Lin, T., Ho, N. & Jordan, M.. (2019). On Efficient Optimal Transport: An Analysis of Greedy and Accelerated Mirror Descent Algorithms. Proceedings of the 36th International Conference on Machine Learning, in PMLR 97:3982-3991
This site last compiled Mon, 16 Sep 2019 16:05:04 +0000
