Optimal Transport Adapter Tuning for Bridging Modality Gaps in Few-Shot Remote Sensing Scene Classification
Zhong JI, Ci LIU, Jingren LIU, Chen TANG, Yanwei PANG, Xuelong LI · 2025
Read Optimal Transport Adapter Tuning for Bridging Modality Gaps in Few-Shot Remote Sensing Scene Classification by Zhong JI, Ci LIU, Jingren LIU, Chen TANG, Yanwei PANG, Xuelong LI (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-2503-14938.txt.
Description
Zhong Ji and co-authors tackle, in a 2025 paper, the task of naming scenes in satellite photos when only a few named samples exist. Photos are rich, but text tags for them are thin, so the two forms do not line up well. The team builds a frame that pulls the two into one shared space, using a math tool that moves mass at least cost. A small add-on part uses cross-form focus to enrich the text side with image facts.
A new loss term weighs hard samples more and keeps the spread of scores in check. That makes the transport step more stable and easier to solve. The frame can grow to new sets with ease. Tests in the paper show it sorts scenes well from just a few samples. The work aids maps, land checks, and quick scans after a storm or quake.
Catalog record
| CREATOR | Zhong JI, Ci LIU, Jingren LIU, Chen TANG, Yanwei PANG, Xuelong LI |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-14938