DisGrem achieves O(epsilon^{-1}) global iteration complexity for driving gradient norm below epsilon in decentralized convex optimization via scheduled consensus and vanishing regularization, matching centralized rates after a burn-in phase.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Distributed Gradient-Regularized Newton Method: Scheduled Consensus and O(epsilon^{-1}) Global Iteration Complexity
DisGrem achieves O(epsilon^{-1}) global iteration complexity for driving gradient norm below epsilon in decentralized convex optimization via scheduled consensus and vanishing regularization, matching centralized rates after a burn-in phase.