Sustainability-aware job scheduler for federated learning

Federated machine learning enables privacy-aware, decentralized training of complex models. Here, we extend the framework DISCO with a task scheduler which distributes work in a sustainable way.