Adaptive Consensus ADMM for Distributed Optimization

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.

ICML 2017

References in corpus (3)

• Training Neural Networks Without Gradients: A Scalable ADMM Approach
• Adaptive Consensus ADMM for Distributed Optimization
• Faster convergence rates of relaxed Peaceman-Rachford and ADMM under regularity assumptions

Cited by in corpus (10)

• Anderson Accelerated Douglas-Rachford Splitting
• Adaptive Consensus ADMM for Distributed Optimization
• Data Multiplexing in Radio Interferometric Calibration
• Distributed ADMM with Synergetic Communication and Computation
• Newton-ADMM: A Distributed GPU-Accelerated Optimizer for Multiclass Classification Problems
• On Consensus-Optimality Trade-offs in Collaborative Deep Learning
• Toward Model Parallelism for Deep Neural Network based on Gradient-free ADMM Framework
• Recycled ADMM: Improve Privacy and Accuracy with Less Computation in Distributed Algorithms
• Generalized Minkowski sets for the regularization of inverse problems
• Divide-and-Conquer MCMC for Multivariate Binary Data