FedMS: Federated Learning with Mixture of Sparsely Activated Foundations Models

TL;DR

FedMS introduces a two-stage federated learning approach with a mixture of foundation models and a sparsely activated LoRA technique, enabling personalized, efficient, and high-performing models in resource-constrained edge scenarios.

Contribution

The paper proposes FedMS, a novel federated learning algorithm combining global and local experts with a mixture model and a sparsely activated LoRA method for efficient personalization.

Findings

FedMS outperforms SOTA baselines by up to 55.25%.

The sparsely activated LoRA improves training efficiency and resource utilization.

The two-stage approach enhances personalization and model performance.

Abstract

Foundation models have shown great success in natural language processing, computer vision, and multimodal tasks. FMs have a large number of model parameters, thus requiring a substantial amount of data to help optimize the model during the training. Federated learning has revolutionized machine learning by enabling collaborative learning from decentralized data while still preserving the data privacy of clients. Despite the great benefits foundation models can have empowered by federated learning, they face severe computation, communication, and statistical challenges. In this paper, we propose a novel two-stage federated learning algorithm called FedMS. A global expert is trained in the first stage and a local expert is trained in the second stage to provide better personalization. We construct a Mixture of Foundation Models (MoFM) with these two experts and design a gate neural network with an inserted gate adapter that joins the aggregation every communication round in the second stage. To further adapt to edge computing scenarios with limited computational resources, we design a novel Sparsely Activated LoRA (SAL) algorithm that freezes the pre-trained foundation model parameters inserts low-rank adaptation matrices into transformer blocks and activates them progressively during the training. We employ extensive experiments to verify the effectiveness of FedMS, results show that FedMS outperforms other SOTA baselines by up to 55.25% in default settings.