FedDHAD weights client models by a learnable estimate of data non-IID-ness and applies adaptive neuron dropout, claiming modest accuracy and speed gains on image classification benchmarks.
Brendan McMahan, Brendan Avent, Aurélien Bellet, and Mehdi Bennis et al
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.DC 1years
2025 1verdicts
REJECT 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Efficient Federated Learning with Heterogeneous Data and Adaptive Dropout
FedDHAD weights client models by a learnable estimate of data non-IID-ness and applies adaptive neuron dropout, claiming modest accuracy and speed gains on image classification benchmarks.