Selective Aggregation for Low-Rank Adaptation in Federated Learning

TL;DR

This paper introduces FedSA-LoRA, a federated learning method that selectively shares low-rank matrices to improve efficiency and performance, based on the distinct roles of matrices in learning general versus client-specific knowledge.

Contribution

The paper proposes FedSA-LoRA, a novel approach that shares only the A matrices in LoRA during federated learning, and extends this paradigm to other LoRA variants, enhancing efficiency and understanding.

Findings

FedSA-LoRA outperforms traditional methods in natural language tasks.

Selective sharing of A matrices maintains model performance while reducing communication.

The approach generalizes well across different LoRA variants.

Abstract

We investigate LoRA in federated learning through the lens of the asymmetry analysis of the learned $A$ and $B$ matrices. In doing so, we uncover that $A$ matrices are responsible for learning general knowledge, while $B$ matrices focus on capturing client-specific knowledge. Based on this finding, we introduce Federated Share-A Low-Rank Adaptation (FedSA-LoRA), which employs two low-rank trainable matrices $A$ and $B$ to model the weight update, but only $A$ matrices are shared with the server for aggregation. Moreover, we delve into the relationship between the learned $A$ and $B$ matrices in other LoRA variants, such as rsLoRA and VeRA, revealing a consistent pattern. Consequently, we extend our FedSA-LoRA method to these LoRA variants, resulting in FedSA-rsLoRA and FedSA-VeRA. In this way, we establish a general paradigm for integrating LoRA with FL, offering guidance for future work on subsequent LoRA variants combined with FL. Extensive experimental results on natural language understanding and generation tasks demonstrate the effectiveness of the proposed method. Our code is available at https://github.com/Pengxin-Guo/FedSA-LoRA.