Spherical Tree-Sliced Wasserstein Distance
Hoang V. TranDepartment of MathematicsNational University of Singapore [email protected], Thanh T. ChuDepartment of Computer Science National University of Singapore [email protected] Khoi N.M. NguyenFPT Software AI Center [email protected] Trang PhamMovian AI, Vietnam [email protected] · 2025
Read Spherical Tree-Sliced Wasserstein Distance by Hoang V. TranDepartment of MathematicsNational University of Singapore [email protected], Thanh T. ChuDepartment of Computer Science National University of Singapore [email protected] Khoi N.M. NguyenFPT Software AI Center [email protected] Trang PhamMovian AI, Vietnam [email protected] (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-2503-11249.txt.
Description
Optimal transport measures how costly it is to reshape one distribution into another, and slicing tricks make that cost cheap to compute in high dimensions. A team of authors extended those tricks in 2025 to data that live on a sphere, such as directions or points on Earth. Their method projects spherical data onto tree-shaped structures on the sphere instead of flat lines, keeping more of the data's shape.
From a new spherical Radon transform they derive closed formulas for the transport cost, giving a distance they call STSW. Theory in the paper shows the distance is well defined and does not depend on how the sphere is rotated. Experiments on gradient flows and self-supervised learning test the metric against recent rivals.
Catalog record
| CREATOR | Hoang V. TranDepartment of MathematicsNational University of Singapore [email protected], Thanh T. ChuDepartment of Computer Science National University of Singapore [email protected] Khoi N.M. NguyenFPT Software AI Center [email protected] Trang PhamMovian AI, Vietnam [email protected] |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 87 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-11249