发表机构
Carnegie Mellon University(卡内基梅隆大学)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
针对 $\kappa$-条件有向图,提出确定性算法,在 $O((\log n + \log^2 \kappa) \log \log n)$ 空间内近似平稳分布,改进了现有随机游走空间界。
AI 中文摘要
对于 $\kappa>1$,一个有向图被称为 $\kappa$-条件图,如果它是 $\kappa$-混合的,并且其平稳分布被均匀分布近似,误差在因子 $\kappa$ 以内。我们提出了一种确定性算法,该算法在 $O((\log n + \log^2 \kappa) \log \log n)$ 空间内,将 $\kappa$-条件图的平稳分布近似到逆多项式相对误差。在 $\kappa = \exp(\Theta(\log^\alpha n))$ 且 $\alpha \in (0, 2/3)$ 的范围内,我们的结果改进了 [Hoza, RANDOM 2021] 中针对一般有向图近似 $\kappa$ 步随机游走的最佳已知 $O(\log n \sqrt{\log \kappa} / \sqrt{\log \log n})$ 空间界。由于最近有关于基于 LLM 的相关问题进展的传闻,以及不确定这些结果何时会出现,我们发布了这个初步版本。进一步的实现细节将在后续版本中提供。
英文摘要
For $κ>1$, a directed graph is $κ$-conditioned if it is $κ$-mixing and its stationary distribution is approximated by the uniform distribution within a factor of $κ$. We present a deterministic algorithm that approximates the stationary distribution of a $κ$-conditioned graph to inverse polynomial relative error in $O((\log n + \log^2 κ) \log \log n)$ space. In the regime $κ= \exp(Θ(\log^αn))$ for any $α\in (0, 2/3)$, our result improves the best-known $O(\log n \sqrt{\log κ} / \sqrt{\log \log n})$ space bound for approximating $κ$-step random walks in general directed graphs by [Hoza, RANDOM 2021]. We release this preliminary version due to recent rumors of LLM-based progress on related problems and uncertainty about when those results may appear. Further implementation details will be provided in a subsequent version.