Toward Data Heterogeneity of Federated Learning

Yuchuan Huang, Chen Hu

Federated learning is a popular paradigm for machine learning. Ideally, federated learning works best when all clients share a similar data distribution. However, it is not always…