Pith. sign in

The grid of ηg is{10−1,10−1/2,10 0,10 1/2,10 1} for Fe- dAvg(M) and SCAFFOLD, and{10−4,10−7/2,10−3,10−5/2,10−2} for FedOPT

Image classification datasetsFor the image classification tasks, we tune ηl over {10−2,10−3/2,10−1,10−1/2,10 0} · 2000

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
browse 1 citing papers

fields

cs.LG 1

years

2025 1

verdicts

CONDITIONAL 1

representative citing papers

FedDuA: Doubly Adaptive Federated Learning

cs.LG · 2025-05-16 · conditional · novelty 6.0

FedDuA sets each round's global learning rate to the average squared client-update norm divided by the aggregated update norm under an adaptive coordinate preconditioner, a rule that is minimax-optimal under an approximate projection condition and convergent for convex objectives.

citing papers explorer

Showing 1 of 1 citing paper.

  • FedDuA: Doubly Adaptive Federated Learning cs.LG · 2025-05-16 · conditional · none · ref 15

    FedDuA sets each round's global learning rate to the average squared client-update norm divided by the aggregated update norm under an adaptive coordinate preconditioner, a rule that is minimax-optimal under an approximate projection condition and convergent for convex objectives.