arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2410.10892stat.MLcs.DScs.LG
收藏

可复现的均匀性测试

Replicable Uniformity Testing

University of California San Diego(加利福尼亚大学圣迭戈分校)

Sihan Liu, Christopher Ye

AI总结:

本文在算法可复现性框架下研究均匀性测试,提出首个在参数ρ上具近线性依赖的可复现测试器,并证明了对称算法近乎匹配的样本复杂度下界。

摘要:

均匀性测试可以说是最基础的分布测试问题之一。给定对未知分布 $\mathbf{p}$ 在 $[n]$ 上的样本访问权限,必须判断 $\mathbf{p}$ 是均匀的,还是在总变差距离上偏离均匀分布 $\varepsilon$。一系列长期的研究确立了均匀性测试的样本复杂度为 $Θ(\sqrt{n}\varepsilon^{-2})$。然而,当输入分布既非均匀也非远离均匀时,已知算法可能表现出高度不可复现的行为。因此,如果将这些算法应用于科学研究,可能会导致相互矛盾的结果,从而削弱公众对科学的信任。在本研究中,我们在算法可复现性框架 [STOC '22] 下重新审视均匀性测试,要求算法在任意分布下均可复现。尽管可复现性通常会在样本复杂度上引入 $ρ^{-2}$ 因子的开销,但我们仅使用 $\tilde{O}(\sqrt{n} \varepsilon^{-2} ρ^{-1})$ 个样本就实现了一个可复现的均匀性测试器。据我们所知,这是首个在 $ρ$ 上具有(近)线性依赖关系的可复现学习算法。最后,我们考虑了一类“对称”算法 [FOCS '00],其输出在域 $[n]$ 重新标记下保持不变,这包括了所有现有的均匀性测试器(包括我们的)。对于这类自然算法,我们证明了可复现均匀性测试的近乎匹配的样本复杂度下界。

英文摘要:

Uniformity testing is arguably one of the most fundamental distribution testing problems. Given sample access to an unknown distribution $\mathbf{p}$ on $[n]$, one must decide if $\mathbf{p}$ is uniform or $\varepsilon$-far from uniform (in total variation distance). A long line of work established that uniformity testing has sample complexity $Θ(\sqrt{n}\varepsilon^{-2})$. However, when the input distribution is neither uniform nor far from uniform, known algorithms may have highly non-replicable behavior. Consequently, if these algorithms are applied in scientific studies, they may lead to contradictory results that erode public trust in science. In this work, we revisit uniformity testing under the framework of algorithmic replicability [STOC '22], requiring the algorithm to be replicable under arbitrary distributions. While replicability typically incurs a $ρ^{-2}$ factor overhead in sample complexity, we obtain a replicable uniformity tester using only $\tilde{O}(\sqrt{n} \varepsilon^{-2} ρ^{-1})$ samples. To our knowledge, this is the first replicable learning algorithm with (nearly) linear dependence on $ρ$. Lastly, we consider a class of ``symmetric" algorithms [FOCS '00] whose outputs are invariant under relabeling of the domain $[n]$, which includes all existing uniformity testers (including ours). For this natural class of algorithms, we prove a nearly matching sample complexity lower bound for replicable uniformity testing.

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。

↑