Communication-Efficient Approximate Gradient Coding for Distributed Learning in Heterogeneous Systems
For distributed learning systems, the work provides a theoretically optimal joint design for gradient coding and quantization to improve straggler resilience and communication efficiency.
The paper proposes a communication-efficient gradient coding scheme that jointly optimizes gradient coding and quantization to address straggler resilience and communication efficiency in heterogeneous distributed learning. Experiments on COCO show significant acceleration in convergence and improved communication efficiency over baselines.
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.