Convergence-Efficient Satellite-Ground Federated Learning for LEO Mega Constellations Optical Networks

We propose an efficient converged satellite-ground federated learning framework by quickly converging parameters to reduce the training time. Simulation results show that it can improve the global aggregation speed.

Saved in:
Bibliographic Details
Main Authors: Ge, Minghao, Zhu, Ruijie, Li, Kai, Wei, Jingbo, Sang, Huiying, Hou, Xiaojie
Format: Conference Proceeding
Language:English
Subjects:
Federated learning
LEO Mega Constellations
Low earth orbit satellites
Optical fiber networks
Satellite constellations
Satellite-ground
Satellites
Simulation
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose an efficient converged satellite-ground federated learning framework by quickly converging parameters to reduce the training time. Simulation results show that it can improve the global aggregation speed.
ISSN:2771-3059
DOI:10.1109/ICOCN59242.2023.10236437