Partial Federated Learning introduces a new algorithm, PartialFL, to train machine learning models with distributed data modalities, improving model performance and addressing data heterogeneity challenges.
Partial Federated Learning (PartialFL) addresses data heterogeneity in federated learning by allowing some data modalities to be shared with the server, leading to improved model performance.