SpaFL: Communication-Efficient Federated Learning with Sparse Models and Low computational Overhead

TL;DR

SpaFL introduces a communication-efficient federated learning framework that employs structured sparsity through trainable thresholds, reducing communication and computation overhead while maintaining high accuracy.

Contribution

It proposes a novel threshold-based pruning method for sparse models in federated learning, optimizing communication by transmitting only thresholds rather than full parameters.

Findings

SpaFL reduces communication costs significantly.

It improves model accuracy compared to sparse baselines.

Theoretical bounds relate sparsity levels to performance.

Abstract

The large communication and computation overhead of federated learning (FL) is one of the main challenges facing its practical deployment over resource-constrained clients and systems. In this work, SpaFL: a communication-efficient FL framework is proposed to optimize sparse model structures with low computational overhead. In SpaFL, a trainable threshold is defined for each filter/neuron to prune its all connected parameters, thereby leading to structured sparsity. To optimize the pruning process itself, only thresholds are communicated between a server and clients instead of parameters, thereby learning how to prune. Further, global thresholds are used to update model parameters by extracting aggregated parameter importance. The generalization bound of SpaFL is also derived, thereby proving key insights on the relation between sparsity and performance. Experimental results show that SpaFL improves accuracy while requiring much less communication and computing resources compared to sparse baselines. The code is available at https://github.com/news-vt/SpaFL_NeruIPS_2024