The GBMA algorithm lets distributed nodes transmit analog gradients over a fading multiple access channel without power control, and provably approaches centralized gradient descent convergence as the number of nodes grows.
Incremental s tochastic subgradient algorithms for convex optimization ,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
On Analog Gradient Descent Learning over Multiple Access Fading Channels
The GBMA algorithm lets distributed nodes transmit analog gradients over a fading multiple access channel without power control, and provably approaches centralized gradient descent convergence as the number of nodes grows.