Dec 6, 2022 · We show that imposing uniformity helps to combat prototype collapse while infusing class semantics improves local models. Extensive experiments ...
People also ask
What is data heterogeneity in federated learning?
Is federated learning a distributed learning?
What is the methodology of federated learning?
What is the use case of federated learning?
We propose FedNH, a novel method that tackles data het- erogeneity with class imbalance by utilizing uniformity and semantics of class prototypes. • We design a ...
We propose FedNH, a novel method that tackles data heterogeneity with class imbalance by utilizing uniformity and semantics of class prototypes.
Feb 7, 2023 · We show that imposing uniformity helps to combat prototype collapse while infusing class semantics improves local models. Extensive experiments ...
This repo provides an implementation of FedNH proposed in for Tackling Data Heterogeneity in Federated Learning with Class Prototypes, which is accepted by ...
Few works address both date heterogeneity and class imbalance without requiring auxiliary datasets or potential privacy leakage. A Motivating Example. Balanced ...
We show that imposing uniformity helps to combat prototype collapse while infusing class semantics improves local models. Extensive experiments were conducted ...
Tackling Data Heterogeneity in Federated Learning with Class Prototypes. Yutong Dai1. Zeyuan Chen2. Junnan Li2. Shelby Heinecke2. Lichao Sun1. Ran Xu2. Lehigh ...
Feb 11, 2023 · FedNH shares the body and the prototypes, with which the attackers may recover the client data of each class. How about the privacy protecting ...
May 27, 2024 · We propose global prototype distillation (FedGPD) for heterogenous federated learning to improve performance of global model.