arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2405.00853cs.LGstat.ML
收藏

图中学习单音半空间的高效算法

Efficient Algorithms for Learning Monophonic Halfspaces in Graphs

Università degli Studi di Milano(米兰大学);Istituto Italiano di Tecnologia(意大利技术研究院);TU Wien(维也纳工业大学)

Marco Bressan, Emmanuel Esposito, Maximilian Thiessen

AI总结:

本文研究图顶点上单音半空间的学习问题,通过结构分析与归约到 2-SAT 给出多项式时间一致性检验,并在监督、在线和主动设置下获得接近最优样本复杂度及相关算法结果。

摘要:

我们研究在图的顶点上学习二分类器的问题。具体而言,我们考虑由单音半空间(monophonic halfspaces)给出的分类器,即对顶点进行划分,且这些划分在某种抽象意义上是凸的。单音半空间以及诸如测地半空间(geodesic halfspaces)等相关概念最近引起了人们的兴趣,并且其性质(例如 VC 维)与底层图 $G$ 的结构之间已建立了若干联系。我们在监督、在线和主动设置下证明了若干关于学习单音半空间的新结果。我们的主要结果是:单音半空间可以在 $n=|V(G)|$ 的多项式时间内以接近最优的被动样本复杂度被学习。这要求我们基于对单音半空间的若干结构性洞见以及归约到 2-可满足性(2-satisfiability),设计一种用于一致性假设检验的多项式时间算法。我们在在线和主动设置下证明了类似结果。我们还表明,该概念类可以以 $\n\operatorname{poly}(n)$ 的延迟被枚举,并且经验风险最小化可以在 $2^{\omega(G)}\operatorname{poly}(n)$ 时间内执行,其中 $\omega(G)$ 是 $G$ 的团数。这些结果回答了文献(González 等人,2020)中的开放问题,并显示出与测地半空间的对比,对于测地半空间,上述部分问题是 NP 难的(Seiffarth 等人,2023)。

英文摘要:

We study the problem of learning a binary classifier on the vertices of a graph. In particular, we consider classifiers given by monophonic halfspaces, partitions of the vertices that are convex in a certain abstract sense. Monophonic halfspaces, and related notions such as geodesic halfspaces,have recently attracted interest, and several connections have been drawn between their properties(e.g., their VC dimension) and the structure of the underlying graph $G$. We prove several novel results for learning monophonic halfspaces in the supervised, online, and active settings. Our main result is that a monophonic halfspace can be learned with near-optimal passive sample complexity in time polynomial in $n = |V(G)|$. This requires us to devise a polynomial-time algorithm for consistent hypothesis checking, based on several structural insights on monophonic halfspaces and on a reduction to $2$-satisfiability. We prove similar results for the online and active settings. We also show that the concept class can be enumerated with delay $\operatorname{poly}(n)$, and that empirical risk minimization can be performed in time $2^{ω(G)}\operatorname{poly}(n)$ where $ω(G)$ is the clique number of $G$. These results answer open questions from the literature (González et al., 2020), and show a contrast with geodesic halfspaces, for which some of the said problems are NP-hard (Seiffarth et al., 2023).

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。

↑