GradMA: A Gradient-Memory-based Accelerated Federated Learning with Alleviated Catastrophic Forgetting

TL;DR

GradMA introduces a novel federated learning method that leverages gradient memory and continual learning principles to mitigate catastrophic forgetting, improve accuracy, and enhance communication efficiency across large-scale, heterogeneous data environments.

Contribution

The paper proposes GradMA, a federated learning approach that uses gradient memory and a memory reduction strategy, with theoretical convergence analysis and superior experimental performance.

Findings

GradMA achieves significant accuracy improvements over SOTA methods.

It reduces communication costs in federated learning.

Theoretical analysis shows linear speed-up with more active workers.

Abstract

Federated Learning (FL) has emerged as a de facto machine learning area and received rapid increasing research interests from the community. However, catastrophic forgetting caused by data heterogeneity and partial participation poses distinctive challenges for FL, which are detrimental to the performance. To tackle the problems, we propose a new FL approach (namely GradMA), which takes inspiration from continual learning to simultaneously correct the server-side and worker-side update directions as well as take full advantage of server's rich computing and memory resources. Furthermore, we elaborate a memory reduction strategy to enable GradMA to accommodate FL with a large scale of workers. We then analyze convergence of GradMA theoretically under the smooth non-convex setting and show that its convergence rate achieves a linear speed up w.r.t the increasing number of sampled active workers. At last, our extensive experiments on various image classification tasks show that GradMA achieves significant performance gains in accuracy and communication efficiency compared to SOTA baselines.