CRAD: Class-wise Reliability-Aware Distillation for Decentralized Heterogeneous Federated Learning
This work proposes Class-wise Reliability-Aware Distillation (CRAD), which, per class, first discards teachers that disagree with the peer consensus and then takes a weighted average of the rest, weighting each teacher by its per-class reliability (precision, or inverse variance).
Baraa Bilbeisi, Meng-Chen Fan, Bao-Cheng Geng et al.
· 1 citation