arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.03380cs.LGcs.ITmath.ITmath.STstat.TH
收藏

相关数据库检测中的相变现象

Phase Transitions in the Detection of Correlated Databases

Ben-Gurion University(本-古里安大学);Tel Aviv University(特拉维夫大学)

Dor Elimelech, Wasim Huleihel

AI总结:

该研究将双高斯数据库相关性检测建模为含未知行排列的假设检验问题,推导了强弱检测的相变阈值,证明了不同渐近条件下检测的不可能性,填补了相关研究空白。

摘要:

我们研究检测两个高斯数据库$\u03a7\in\mathbb{R}^{n\times d}$与$\u03a9^{n\times d}$之间相关性的问题,每个数据库均由$n$个用户、$d$个特征构成。该问题与社交媒体分析、计算生物学等领域相关。我们将其建模为假设检验问题:零假设下,两个数据库统计独立;备择假设下,存在$n$个用户集合上的未知排列$\u03c3$(即行排列),使得$\u03a7$与$\u03a9^\u03c3$($\u03a9$的排列版本)呈$\u03c1$相关。我们确定了最优检验发生相变的精确阈值,阈值依赖于$n$与$d$的渐近 regime。具体而言,我们证明若当$d\to\infty$时$\u03c1^2d\to0$,则无论$n$取何值,弱检测(性能略优于随机猜测)在统计上不可能实现,这补充了一种对$\u03a7^T\u03a9$所有元素求和取阈值的简单检验的性能结论。此外,当$d$固定时,我们证明对任意$\u03c1<\u03c1^\star$($\u03c1^\star$是$d$的显式函数),强检测(误差概率趋于零)不可能实现,且只要$\u03c1^2d\to0$,弱检测同样不可能实现。这些结果填补了近期相关研究中的重要空白。

英文摘要:

We study the problem of detecting the correlation between two Gaussian databases $\mathsf{X}\in\mathbb{R}^{n\times d}$ and $\mathsf{Y}^{n\times d}$, each composed of $n$ users with $d$ features. This problem is relevant in the analysis of social media, computational biology, etc. We formulate this as a hypothesis testing problem: under the null hypothesis, these two databases are statistically independent. Under the alternative, however, there exists an unknown permutation $σ$ over the set of $n$ users (or, row permutation), such that $\mathsf{X}$ is $ρ$-correlated with $\mathsf{Y}^σ$, a permuted version of $\mathsf{Y}$. We determine sharp thresholds at which optimal testing exhibits a phase transition, depending on the asymptotic regime of $n$ and $d$. Specifically, we prove that if $ρ^2d\to0$, as $d\to\infty$, then weak detection (performing slightly better than random guessing) is statistically impossible, irrespectively of the value of $n$. This compliments the performance of a simple test that thresholds the sum all entries of $\mathsf{X}^T\mathsf{Y}$. Furthermore, when $d$ is fixed, we prove that strong detection (vanishing error probability) is impossible for any $ρ<ρ^\star$, where $ρ^\star$ is an explicit function of $d$, while weak detection is again impossible as long as $ρ^2d\to0$. These results close significant gaps in current recent related studies.

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。

↑