arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.01068cs.LG
收藏

FedLAP-DP:通过共享差分隐私损失近似实现联邦学习

FedLAP-DP: Federated Learning by Sharing Differentially Private Loss Approximations

CISPA Helmholtz Center for Information Security(CISPA亥姆霍兹信息安全中心)

Hui-Po Wang, Dingfan Chen, Raouf Kerkouche, Mario Fritz

AI总结:

本文提出 FedLAP-DP,通过共享模拟局部梯度的差分隐私合成损失样本近似全局损失景观,在相同隐私成本下提升非独立同分布数据上的隐私-效用权衡与收敛速度。

摘要:

传统的联邦学习(FL)梯度共享方法(如 FedAvg)依赖于局部模型的聚合,在差分隐私(DP)机制或数据异质性下往往面临性能下降,这可归因于局部目标与全局目标之间的不一致。为解决这一问题,我们提出了 FedLAP-DP,一种新颖的联邦学习隐私保护方法。在我们的形式化中,客户端通过模拟局部区域内真实图像的梯度,合成一小组样本来近似局部损失景观。这些合成样本作为损失替代物,在服务器端进行聚合,以揭示全局损失景观并实现全局优化。基于这些洞见,我们为在联邦学习中实施记录级差分隐私提供了新视角。正式的隐私分析表明,FedLAP-DP 产生与典型梯度共享方案相同的隐私成本,同时在隐私与效用之间实现了更优的权衡。大量实验验证了我们的方法在 DP 和非 DP 设置下,于具有高度偏斜分布的各类数据集上的优越性。除了可观的性能外,与典型梯度共享方法相比,我们的方法具有更快的收敛速度,并通过发送更多合成图像,为以通信成本换取更好性能开辟了可能性。源代码可在 https://github.com/hui-po-wang/FedLAP-DP 获取。

英文摘要:

Conventional gradient-sharing approaches for federated learning (FL), such as FedAvg, rely on aggregation of local models and often face performance degradation under differential privacy (DP) mechanisms or data heterogeneity, which can be attributed to the inconsistency between the local and global objectives. To address this issue, we propose FedLAP-DP, a novel privacy-preserving approach for FL. Our formulation involves clients synthesizing a small set of samples that approximate local loss landscapes by simulating the gradients of real images within a local region. Acting as loss surrogates, these synthetic samples are aggregated on the server side to uncover the global loss landscape and enable global optimization. Building upon these insights, we offer a new perspective to enforce record-level differential privacy in FL. A formal privacy analysis demonstrates that FedLAP-DP incurs the same privacy costs as typical gradient-sharing schemes while achieving an improved trade-off between privacy and utility. Extensive experiments validate the superiority of our approach across various datasets with highly skewed distributions in both DP and non-DP settings. Beyond the promising performance, our approach presents a faster convergence speed compared to typical gradient-sharing methods and opens up the possibility of trading communication costs for better performance by sending a larger set of synthetic images. The source is available at \url{https://github.com/hui-po-wang/FedLAP-DP}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑