Data Deduplication

This paper introduces FEDDE, a general and efficient framework that addresses data redundancy across clients to facilitate effective federated learning (FL). At its core, FEDDE adopts a hierarchical deduplication architecture where clients first …