Adaptive Consensus ADMM for Distributed Optimization
Machine Learning
2017-06-21 v2 Numerical Analysis
Systems and Control
Abstract
The alternating direction method of multipliers (ADMM) is commonly used for distributed model fitting problems, but its performance and reliability depend strongly on user-defined penalty parameters. We study distributed ADMM methods that boost performance by using different fine-tuned algorithm parameters on each worker node. We present a O(1/k) convergence rate for adaptive ADMM methods with node-specific parameters, and propose adaptive consensus ADMM (ACADMM), which automatically tunes parameters without user oversight.
Cite
@article{arxiv.1706.02869,
title = {Adaptive Consensus ADMM for Distributed Optimization},
author = {Zheng Xu and Gavin Taylor and Hao Li and Mario Figueiredo and Xiaoming Yuan and Tom Goldstein},
journal= {arXiv preprint arXiv:1706.02869},
year = {2017}
}
Comments
ICML 2017