Decentralized SGD achieves high-probability convergence with order-optimal rates and linear speedup under standard cost assumptions matching those for MSE convergence.
Diffusion Least-Mean Squares Over Adaptive Net- works: Formulation and Performance Analysis,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.LG 2verdicts
UNVERDICTED 2representative citing papers
GT-DSGD achieves order-optimal high-probability rates O(log(1/δ)/sqrt(nT)) for non-convex and O(log(1/δ)/(nT)) for PL costs, matching the conditions used for mean-squared error bounds.
citing papers explorer
-
High-Probability Convergence Guarantees of Decentralized SGD
Decentralized SGD achieves high-probability convergence with order-optimal rates and linear speedup under standard cost assumptions matching those for MSE convergence.
-
High-Probability Convergence in Decentralized Stochastic Optimization with Gradient Tracking
GT-DSGD achieves order-optimal high-probability rates O(log(1/δ)/sqrt(nT)) for non-convex and O(log(1/δ)/(nT)) for PL costs, matching the conditions used for mean-squared error bounds.