Update-based estimates of client data heterogeneity in sub-model federated learning are dominated by device capacity, and adaptive allocation adds nothing over a matched-budget random control once parameter coverage is guaranteed.
On the convergence and stability of distributed sub- model training,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Capacity Confounds and Coverage Guarantees in Adaptive Sub-model Federated Learning
Update-based estimates of client data heterogeneity in sub-model federated learning are dominated by device capacity, and adaptive allocation adds nothing over a matched-budget random control once parameter coverage is guaranteed.