ATRS uses a shared neural policy in a multi-agent MDP to adaptively re-split trajectory segments during parallel ADMM optimization, cutting iterations by up to 26% and time by 19.1% with zero-shot generalization.
Distributed optimization and statistical learning via the alternating direction method of multipliers,
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
The Dual² approach produces iD2A and MiD2A gradient methods that achieve asymptotic convergence under milder conditions on the public function and linear rates with reduced communication and computation complexity.
citing papers explorer
-
ATRS: Adaptive Trajectory Re-splitting via a Shared Neural Policy for Parallel Optimization
ATRS uses a shared neural policy in a multi-agent MDP to adaptively re-split trajectory segments during parallel ADMM optimization, cutting iterations by up to 26% and time by 19.1% with zero-shot generalization.
-
Accelerated Decentralized Constraint-Coupled Optimization: A Dual$^2$ Approach
The Dual² approach produces iD2A and MiD2A gradient methods that achieve asymptotic convergence under milder conditions on the public function and linear rates with reduced communication and computation complexity.