English(EN) Federated Differentially Private Optimization via Low-Rank Preconditioning

新框架降低了联邦隐私优化的通信成本

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-23 04:00

研究人员开发了一种用于联邦学习环境中差分隐私（DP）优化新框架。该方法称为DP-PCA，旨在降低通信成本，同时保持中央DP保证。它通过DP秩选择步骤、DP幂法以及投影协方差矩阵的噪声释放来实现，能够以比现有集中式方法显著更低的通信量实现预处理DP梯度下降。 AI

影响这项研究可以实现跨分布式数据集的机器学习模型更高效和私密的训练。

排序理由该集群包含一篇研究论文，详细介绍了联邦学习中差分隐私优化的一种新方法。[lever_c_demoted from research: ic=1 ai=1.0]

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 4 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇研究论文，详细介绍了联邦学习中差分隐私优化的一种新方法。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, safety; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

arXiv cs.LG TIER_1 English(EN) · Antti Koskela, Tejas Kulkarni, Elias Raninen · 2026-10-23 04:00

Federated Differentially Private Optimization via Low-Rank Preconditioning

arXiv:2610.31069v1 Announce Type: new Abstract: We consider differentially private (DP) optimization in the federated setting, where a DP estimate of the feature covariance matrix serves as a Hessian approximation for generalized linear models. Existing central DP methods for cov…