Communication-Efficient Approximate Gradient Coding for Distributed Learning in Heterogeneous Systems

TL;DR

This paper introduces a communication-efficient gradient coding scheme for distributed learning that improves convergence speed and reduces communication costs in heterogeneous systems.

Contribution

It presents a unified framework optimizing gradient coding and quantization, with a closed-form code structure and a low-complexity bit allocation algorithm.

Findings

Accelerates convergence on COCO dataset

Enhances communication efficiency over baselines

Provides rigorous convergence analysis for convex functions

Abstract

We propose a communication-efficient optimally structured gradient coding scheme to jointly address straggler resilience and communication efficiency in heterogeneous distributed learning. By establishing a unified framework that simultaneously optimizes gradient coding and quantization, we formulate an optimization problem to minimize residual error subject to an unbiasedness constraint. We rigorously establish the joint global optimum by deriving a closed-form code structure coupled with an optimal bit allocation strategy, while simultaneously proposing a low-complexity bit allocation algorithm that efficiently yields near-optimal performance. We provide rigorous convergence analysis for convex and smooth functions. Experiments on the COCO dataset demonstrate that our joint design significantly accelerates convergence and enhances communication efficiency compared to existing baselines.