Coded Stochastic ADMM for Decentralized Consensus Optimization with Edge Computing

TL;DR

This paper introduces a coded stochastic ADMM algorithm for decentralized consensus optimization in edge computing environments, effectively addressing communication bottlenecks and slow nodes, with proven convergence and robust performance.

Contribution

It develops a novel coded stochastic ADMM method that improves communication efficiency and robustness in decentralized edge computing systems for large-scale machine learning.

Findings

Converges at a rate of O(1/√k) with appropriate mini-batch size.

Outperforms existing algorithms in communication efficiency and response speed.

Robust against straggler nodes in distributed networks.

Abstract

Big data, including applications with high security requirements, are often collected and stored on multiple heterogeneous devices, such as mobile devices, drones and vehicles. Due to the limitations of communication costs and security requirements, it is of paramount importance to extract information in a decentralized manner instead of aggregating data to a fusion center. To train large-scale machine learning models, edge/fog computing is often leveraged as an alternative to centralized learning. We consider the problem of learning model parameters in a multi-agent system with data locally processed via distributed edge nodes. A class of mini-batch stochastic alternating direction method of multipliers (ADMM) algorithms is explored to develop the distributed learning model. To address two main critical challenges in distributed networks, i.e., communication bottleneck and straggler nodes (nodes with slow responses), error-control-coding based stochastic incremental ADMM is investigated. Given an appropriate mini-batch size, we show that the mini-batch stochastic ADMM based method converges in a rate of $O(\frac{1}{\sqrt{k}})$, where $k$ denotes the number of iterations. Through numerical experiments, it is revealed that the proposed algorithm is communication-efficient, rapidly responding and robust in the presence of straggler nodes compared with state of the art algorithms.