Joint Metric Space Embedding by Unbalanced OT with Gromov–-Wasserstein Marginal Penalization
Florian Beier, Moritz Piening, Robert Beinert, Gabriele Steidl · 2025
Read Joint Metric Space Embedding by Unbalanced OT with Gromov–-Wasserstein Marginal Penalization by Florian Beier, Moritz Piening, Robert Beinert, Gabriele Steidl (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-07510.txt.
Description
Florian Beier, Moritz Piening, Robert Beinert, and Gabriele Steidl present a way to align two very different data sets into one shared space, in a research paper from 2025. The data sets come from different domains and share no known matches. Their method is based on optimal transport, which moves mass between distributions at least cost.
The twist is a penalty that keeps the internal distances within each data set sensible. The authors prove that a best solution exists and that it approaches a known distance measure as the penalty grows. They also turn the problem into a form a computer can solve step by step. Examples cover both flat and curved shared spaces.
Catalog record
| CREATOR | Florian Beier, Moritz Piening, Robert Beinert, Gabriele Steidl |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-07510