FedX: Explanation-Guided Pruning for Communication-Efficient Federated Learning in Remote Sensing
Barış Büyüktaş, Jonas Klotz, Begüm Demir · 2025
Read FedX: Explanation-Guided Pruning for Communication-Efficient Federated Learning in Remote Sensing by Barış Büyüktaş, Jonas Klotz, Begüm Demir (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-2508-06256.txt.
Description
Barış Büyüktaş, Jonas Klotz, and Begüm Demir cut shared model size in joint learning with guided pruning, in work from 2025. Joint learning trains one model across split data stores that cannot pool raw data. Earth viewing photos often face legal and privacy bars on central storage, so joint learning fits well.
Big model updates sent each round drive up link cost for remote sites. Their method, called FedX, scores part importance with explanation tools at the server. Least useful parts are pruned before the slim model goes back to clients. Tests on two scene sorting datasets beat plain and rival pruning runs. Shared parameter counts fell a lot while global scores rose.
Catalog record
| CREATOR | Barış Büyüktaş, Jonas Klotz, Begüm Demir |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-06256