Neural Optimal Transport

Alexander Korotin, Daniil Selikhanovych, Evgeny Burnaev
1/28/2022
cs.LG

Official code: iamalexkorotin/NeuralOptimalTransport (239 stars, MIT, last commit 2026-05-04).

Abstract

We present a novel neural-networks-based algorithm to compute optimal transport maps and plans for strong and weak transport costs. To justify the usage of neural networks, we prove that they are universal approximators of transport plans between probability distributions. We evaluate the performance of our optimal transport algorithm on toy examples and on the unpaired image-to-image translation.

View on arXivView PDF

Code Implementations(1)

iamalexkorotin/NeuralOptimalTransportOfficial100%
MIT

PyTorch implementation of "Neural Optimal Transport" (ICLR 2023 Spotlight)

23926Python, Jupyter NotebookFeb 4, 20235 months agoMIT
deep-learningimage-to-image-translationneural-networksneural-optimal-transportoptimal-transport+1 more

Similar Papers

Do Neural Optimal Transport Solvers Work? A Continuous Wasserstein-2 Benchmark

86% match2 repos672021

Improving Neural Optimal Transport via Displacement Interpolation

86% match2024

OT-Net: A Reusable Neural Optimal Transport Solver

85% match2023

Computing high-dimensional optimal transport by flow neural networks

84% match1 repo22023

Neural Conditional Transport Maps

84% match2025

Embedding Empirical Distributions for Computing Optimal Transport Maps

83% match1 repo42025

Cite this paper

@article{korotin2022neural,
  title  = {Neural Optimal Transport},
  author = {Alexander Korotin and Daniil Selikhanovych and Evgeny Burnaev},
  year   = {2022},
  eprint = {2201.12220},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2201.12220}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2201.12220.svg)](https://codeofpaper.com/paper/2201.12220)

Discussion