在设备端大模型微调中，用差分隐私保护用户数据隐私。
DP-FedLoRA: Privacy-Enhanced Federated Fine-Tuning for On-Device Large Language Models
- 客户端对低秩适配矩阵加高斯噪声，实现差分隐私。
- 在多个基准上性能接近非隐私版本，且隐私预算可调。
- 适合注重隐私的边缘智能场景，如手机、IoT设备。
随着设备端大语言模型系统日益普及，联邦微调使语言理解与生成能力直接在边缘设备上实现；然而，其涉及处理敏感的用户特定数据，在联邦学习框架下引发显著隐私担忧。为应对这一挑战，我们提出 DP-FedLoRA，一种结合基于 LoRA 的适配与差分隐私的通信高效联邦微调框架。每个客户端本地对 LoRA 矩阵进行裁剪并添加高斯噪声，以满足 (ε, δ)-差分隐私。我们进一步提供理论分析，证明更新无偏性，并推导出噪声引入方差的边界，为隐私预算校准提供实用指导。跨主流基准的实验结果表明，DP-FedLoRA 在保持强隐私保障的同时，仍具备竞争性性能，为设备端 LLM 的可扩展、隐私保护部署铺平道路。
原文摘要 · Abstract (English)
As on-device large language model (LLM) systems become increasingly prevalent, federated fine-tuning enables advanced language understanding and generation directly on edge devices; however, it also involves processing sensitive, user-specific data, raising significant privacy concerns within the federated learning framework. To address these challenges, we propose DP-FedLoRA, a privacy-enhanced federated fine-tuning framework that integrates LoRA-based adaptation with differential privacy in a communication-efficient setting. Each client locally clips and perturbs its LoRA matrices using Gaussian noise to satisfy ($ε$, $δ$)-differential privacy. We further provide a theoretical analysis demonstrating the unbiased nature of the updates and deriving bounds on the variance introduced by noise, offering practical guidance for privacy-budget calibration. Experimental results across mainstream benchmarks show that DP-FedLoRA delivers competitive performance while offering strong privacy guarantees, paving the way for scalable and privacy-preserving LLM deployment in on-device environments.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。