---
格式版本: 2
标题: "Distributed Online Estimation of Spiked Eigenvalues with Adaptive Weighting under Persistent Aspect Ratio Heterogeneity"
原文链接: "https://arxiv.org/abs/2608.19045"
发布日期: "2026-08-19"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Wed, 19 Aug 2026 15:38:05 UTC (2,727 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T14:41:10+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T14:38:42+08:00"
入库时间: "2026-08-20T06:41:10.447Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=Oracle&searchtype=all"
匹配关键词:
  []
相关厂家:
  - "Oracle"
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 11
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "内容为数学统计理论论文，涉及稀疏协方差特征值在线估计，与超节点/AI Rack/机柜级AI基础设施完全无关，无任何相关技术、商业或部署信息。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T14:42:25+08:00"
AI主题相关性: 0
AI来源权威性: 10
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 1
AI摘要: "针对各节点有效样本量持续异质的高维场景，本文提出一种“先校正再聚合”的分布式在线尖峰特征值估计框架，每个节点通过逆瑞利变换消除偏差，服务器用自适应 soft-max 权重融合，每轮仅传输 O(k) 个标量。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:19:42.419Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.19045"
---

## Mathematics > Statistics Theory

## Title:Distributed Online Estimation of Spiked Eigenvalues with Adaptive Weighting under Persistent Aspect Ratio Heterogeneity

Authors:[Lu Yan](https://arxiv.org/search/math?searchtype=author&query=Yan,+L), [Jiang Hu](https://arxiv.org/search/math?searchtype=author&query=Hu,+J), [Yonghan Zhang](https://arxiv.org/search/math?searchtype=author&query=Zhang,+Y), [Xiaoyue Li](https://arxiv.org/search/math?searchtype=author&query=Li,+X)

[View PDF](https://arxiv.org/pdf/2608.19045) [HTML (experimental)](https://arxiv.org/html/2608.19045v1)

> Abstract:We study online estimation of spiked covariance eigenvalues from observations distributed across $L$ nodes with heterogeneous and persistent effective sample sizes. In the proportional high-dimensional regime, local Rayleigh statistics are deterministically distorted by node-specific aspect ratios $c_{\ell,t}=p/N^{\mathrm{eff}}_{\ell,t}$, and direct aggregation of uncorrected statistics converges to the wrong limit. We propose a correct-then-aggregate framework in which each node removes its deterministic bias via an inverse Rayleigh transfer map, and the server fuses corrected estimates using adaptive soft-max weights based on predictable fluctuation metrics, transmitting only $O(k)$ scalars per active node per round. We establish consistency and asymptotic normality of the global estimator, enabling valid online inference, and derive non-asymptotic bounds quantifying how accuracy improves with the number of nodes and their effective sample sizes. The adaptive weights achieve variance reduction comparable to oracle inverse-variance weighting, confirming the data-driven construction is nearly efficient. Simulation studies validate these properties. An application to cross-venue monitoring of a dominant market factor shows the method tracks systemic risk in real time while substantially reducing communication cost relative to a centralized pooled approach.

| Subjects: | Statistics Theory (math.ST) |
| --- | --- |
| Cite as: | [arXiv:2608.19045](https://arxiv.org/abs/2608.19045) \[math.ST\] |
|  | (or [arXiv:2608.19045v1](https://arxiv.org/abs/2608.19045v1) \[math.ST\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.19045](https://doi.org/10.48550/arXiv.2608.19045) |

## Submission history

From: Lu Yan \[[view email](https://arxiv.org/show-email/50772740/2608.19045)\]  
**\[v1\]** Wed, 19 Aug 2026 15:38:05 UTC (2,727 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.19045) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))