OTCOP: Learning optimal transport maps via constraint optimizations

Xiaokai Huo, Hailiang Liu

Keywords: optimal transport, constraint optimization, Monge problem

TL;DR: We integrates constraint optimization algorithm and neural networks for the computation of optimal transport maps based on the Monge formulation.

Abstract: The approximation power of neural networks makes it an ideal tool to learn optimal transport maps. However, existing methods are mostly based on the Kantorovich duality and require regularization and/or special network structures such as Input Convex Neural Networks (ICNN). In this paper, we propose a direct constraint optimization algorithm for the computation of optimal transport maps based on the Monge formulation. We solve this constraint optimization problem by using three different methods: the penalty method, the augmented Lagrangian method, and the alternating direction method of multipliers method (AMDD). We demonstrate a significant improvement in the accuracy of learned optimal transport maps on benchmarks. Moreover, we show that our methods reduce the regularization effects and accurately learn the target distributions at lower transport cost.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning