Data-Free Knowledge Distillation for Heterogeneous Federated Learning

                   Zhuangdi Zhu 1 Junyuan Hong 1 Jiayu Zhou 1

             Abstract                privacy-preserving learning scheme, FL has shown its po-
    Federated Learning (FL) is a decentralized         tential to facilitate real-world applications, including health-
    machine-learning paradigm in which a global        care (Sheller et al., 2020), biometrics (Aggarwal et al.,
    server iteratively aggregates the m ...