On Unbalanced Optimal Transport: Gradient Methods, Sparsity and Approximation Error

We study the Unbalanced Optimal Transport (UOT) between two measures of possibly different masses with at most $n$ components, where the marginal constraints of standard Optimal Transport (OT) are relaxed via Kullback-Leibler divergence with regularization factor $τ$.

Open

Year: 2022
ArXiv: arxiv.org/abs/2202.03618
URL: arxiv.org/abs/2202.03618v4
Hosting: External sourcelicense unknown

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text: arxiv.org/abs/2202.03618v4
TL;DR: Semantic Scholar

Attribution policy →