Back to Search Start Over

Promoting Data and Model Privacy in Federated Learning through Quantized LoRA

Authors :
Zhu, JianHao
Lv, Changze
Wang, Xiaohua
Wu, Muling
Liu, Wenhao
Li, Tianlong
Ling, Zixuan
Zhang, Cenyuan
Zheng, Xiaoqing
Huang, Xuanjing
Publication Year :
2024

Abstract

Conventional federated learning primarily aims to secure the privacy of data distributed across multiple edge devices, with the global model dispatched to edge devices for parameter updates during the learning process. However, the development of large language models (LLMs) requires substantial data and computational resources, rendering them valuable intellectual properties for their developers and owners. To establish a mechanism that protects both data and model privacy in a federated learning context, we introduce a method that just needs to distribute a quantized version of the model's parameters during training. This method enables accurate gradient estimations for parameter updates while preventing clients from accessing a model whose performance is comparable to the centrally hosted one. Moreover, we combine this quantization strategy with LoRA, a popular and parameter-efficient fine-tuning method, to significantly reduce communication costs in federated learning. The proposed framework, named \textsc{FedLPP}, successfully ensures both data and model privacy in the federated learning context. Additionally, the learned central model exhibits good generalization and can be trained in a resource-efficient manner.

Details

Database :
arXiv
Publication Type :
Report
Accession number :
edsarx.2406.10976
Document Type :
Working Paper