New DP-μP framework enables hyperparameter transfer for private model training

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-19 04:00

研究人员开发了DP-μP，一个将最大更新参数化（μP）扩展到差分隐私优化中的新框架。标准的μP在正常训练中允许零样本超参数迁移，但在DP约束下，由于噪声注入和梯度裁剪而失效。DP-μP通过分析这些隐私机制与特征学习之间的相互作用来解决这个问题，推导出了层级学习率缩放的信噪比（SNR）条件。该框架为DP-SGD、DP-Adam和DP-Muon优化器提供了具体的缩放规则，能够有效地在不同模型宽度和MLP、ViT、GPT-2等架构之间进行零样本学习率迁移，同时降低了计算成本。 AI

影响在差分隐私下实现更高效的大规模模型训练，降低计算成本并提高性能。

排序理由该集群描述了一个新的理论框架和经验验证，用于差分隐私优化中的超参数迁移，该内容发表在arXiv论文中。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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

New DP-μP framework enables hyperparameter transfer for private model training

本文如何被排名

Signal score: 10 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群描述了一个新的理论框架和经验验证，用于差分隐私优化中的超参数迁移，该内容发表在arXiv论文中。[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, model release; 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。

报道来源 [1]

arXiv cs.LG TIER_1 English(EN) · Tianze Wang, Zhiqi Bu, Linjun Zhang · 2026-10-19 04:00

迈向差分隐私优化中的超参数迁移

arXiv:2610.22976v1 Announce Type: new Abstract: Scaling laws and hyperparameter tuning remain central challenges in training large-scale models, particularly under differential privacy (DP), where optimization is further complicated by noise injection and per-sample gradient clip…

本文如何被排名

报道来源 [1]

迈向差分隐私优化中的超参数迁移

相关实体

相关话题