Ground-Assisted Federated Learning in LEO Satellite Constellations
This addresses the challenge of distributed training in space-based applications, offering a robust solution for satellite imaging inference, though it appears incremental as it builds on FedAvg.
The paper tackles the problem of collaborative machine learning in LEO satellite constellations without sharing local datasets, proposing a novel asynchronous federated learning algorithm that shows fast convergence and excellent test accuracy on MNIST and CIFAR-10 datasets.
In Low Earth Orbit (LEO) mega constellations, there are relevant use cases, such as inference based on satellite imaging, in which a large number of satellites collaboratively train a machine learning model without sharing their local datasets. To address this problem, we propose a new set of algorithms based on Federated learning (FL), including a novel asynchronous FL procedure based on FedAvg that exhibits better robustness against heterogeneous scenarios than the state-of-the-art. Extensive numerical evaluations based on MNIST and CIFAR-10 datasets highlight the fast convergence speed and excellent asymptotic test accuracy of the proposed method.