Reproduce “Smooth and Sparse Optimal Transport” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Mathieu Blondel, Vivien Seguy, Antoine Rolet

The authors introduce a smooth optimal-transport method using a squared 2-norm penalty instead of entropy regularization. Its dual and semi-dual formulations can be optimized with gradient-based methods, while the plan sets entries below a cost-dependent threshold to zero. In color transfer with 4,096 centroids per image, a plan has 99% zero entries, versus 0% for entropy; this is not a runtime comparison.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.