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 1911.07976cs.ITcs.DScs.LGmath.IT
收藏

常数空间中分布的熵估计

Estimating Entropy of Distributions in Constant Space

Cornell University(康奈尔大学);Massachusetts Institute of Technology(麻省理工学院)

Jayadev Acharya, Sourbh Bhadane, Piotr Indyk, Ziteng Sun

摘要:

我们考虑的任务,估计熵的$k$ ary分布的样本中的流模型,其中空间是有限的。我们的主要贡献是一个算法,需要$O\left(\frac{k \log (1/\varepsilon)^2}{\varepsilon^3}\right)$样本和一个恒定的$O(1)$内存字的空间和输出的$H(p)$的$\pm\varepsilon$估计。在没有空间限制的情况下,样本复杂度已经被建立为$S(k,\varepsilon)=Θ\left(\frac k{\varepsilon\log k}+\frac{\log^2 k}{\varepsilon^2}\right)$,其在域大小$k$中是次线性的,并且实现最佳样本复杂度的当前算法也需要$k$中的近线性空间。我们的算法划分$[0,1]$到区间,并估计在每个区间的概率值的熵贡献。区间设计用于权衡这些估计值的偏差和方差。

英文摘要:

We consider the task of estimating the entropy of $k$-ary distributions from samples in the streaming model, where space is limited. Our main contribution is an algorithm that requires $O\left(\frac{k \log (1/\varepsilon)^2}{\varepsilon^3}\right)$ samples and a constant $O(1)$ memory words of space and outputs a $\pm\varepsilon$ estimate of $H(p)$. Without space limitations, the sample complexity has been established as $S(k,\varepsilon)=Θ\left(\frac k{\varepsilon\log k}+\frac{\log^2 k}{\varepsilon^2}\right)$, which is sub-linear in the domain size $k$, and the current algorithms that achieve optimal sample complexity also require nearly-linear space in $k$. Our algorithm partitions $[0,1]$ into intervals and estimates the entropy contribution of probability values in each interval. The intervals are designed to trade off the bias and variance of these estimates.

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。

↑