FedAvg matches centralized training accuracy on mammography data split by breast density heterogeneity, showing standard FL can handle this clinical variation without special fixes.
Cancers , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Evaluating Federated Learning approaches for mammography under breast density heterogeneity
FedAvg matches centralized training accuracy on mammography data split by breast density heterogeneity, showing standard FL can handle this clinical variation without special fixes.