arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.11547math.STmath.PRstat.MLstat.OTstat.TH
收藏

过渡条件独立性

Transitional Conditional Independence

AI4Science Lab Korteweg-de Vries Institute for Mathematics University of Amsterdam(阿姆斯特丹大学柯特韦格-德弗里斯数学研究所AI4科学实验室)

Patrick Forré

AI总结:

本文提出过渡条件独立性,针对含非随机变量的统计模型,证明其非对称性必要,推导分离规则并公理化结构,应用于辅助性、不变预测等场景,解决普通条件独立性的局限。

摘要:

统计模型包含非随机变量:参数、处理、环境、设计点。普通条件独立性无法表达涉及此类变量的关系,应用它必须先为这些变量赋予分布,这会改变陈述的含义。本文引入过渡条件独立性,它关联马尔可夫核$K(W|T)$上的三个变量,其中$T$为非随机输入,定义为单一分解式:$X\boldsymbol{\rotatebox[origin=c]{270}{$\not\rotatebox[origin=c]{90}{$\rotatebox[origin=c]{90}{$\boldsymbol{\bot}}$}}}\boldsymbol{\rotatebox[origin=c]{270}{$\not\rotatebox[origin=c]{90}{$\bot}}}_{K(W|T)} Y |Z \iff \exists\\, Q(X|Z):\\; K(X,Y,Z|T) = Q(X|Z)\otimes K(Y,Z|T)$。该关系断言存在对每个输入$t$均相同的马尔可夫核$Q(X|Z)$,因此得到分解式而非条件期望间的几乎必然恒等式,且无需在输入空间上赋予分布。该关系是非对称的,我们证明这种非对称性是必要的:将其对称化会破坏其原本要表达的陈述。我们证明除对称性外的所有分离规则的左、右版本,其中10条适用于任意可测空间,其余需满足所涉及空间的一个条件,并给出对称性本身成立的判据。我们将所得结构公理化,证明其源自任意对称分离体的移位,并给出若干应用:辅助性、充分性和充分性成为在参数上逐点成立的分解式,无需先验和零集;Fisher-Neyman定理和Basu定理取此形式;不变预测的不变性假设$Y \boldsymbol{\rotatebox[origin=c]{270}{$\not\rotatebox[origin=c]{90}{$\rotatebox[origin=c]{90}{$\boldsymbol{\bot}}$}}}\boldsymbol{\rotatebox[origin=c]{270}{$\not\rotatebox[origin=c]{90}{$\bot}}} E | X_S$获得其预期含义:存在一个核在每个环境$E$中从$X_S$预测$Y$;具有非随机输入节点的贝叶斯网络满足有向全局马尔可夫性质,其图形d-分离判据返回马尔可夫核的分解式,适用于任意输入空间。

英文摘要:

Statistical models contain variables that are not random: parameters, treatments, environments, design points. Ordinary conditional independence cannot express relations involving such variables. To apply it one must first put a distribution on them, and that changes the meaning of the statement. This paper introduces transitional conditional independence. It relates three variables on a Markov kernel $K(W|T)$ with non-stochastic input $T$, and is defined by a single factorization: \[ X\perp\!\!\perp_{K(W|T)} Y |Z \quad :\iff \quad \exists\, Q(X|Z):\; K(X,Y,Z|T) = Q(X|Z)\otimes K(Y,Z|T).\] The relation asserts a Markov kernel $Q(X|Z)$ that is the same for every input $t$. It therefore yields a factorization rather than an almost-sure identity between conditional expectations, and it needs no distribution on the input space. The relation is asymmetric. We show that the asymmetry is essential: symmetrizing it destroys the statements it was built to make. We prove left and right versions of all separoid rules except Symmetry. Ten of them hold on arbitrary measurable spaces, the remaining ones under one condition on the spaces involved, and we give criteria for when Symmetry itself holds. We axiomatize the resulting structure and show that it arises from any symmetric separoid by a shift. We give several applications. Ancillarity, sufficiency and adequacy become factorizations that hold pointwise in the parameter, without a prior and without null sets; the theorems of Fisher--Neyman and of Basu take this form. The invariance hypothesis of invariant prediction, $Y \perp\!\!\perp E | X_S$, receives its intended meaning: one kernel predicts $Y$ from $X_S$ in every environment $E$. And Bayesian networks with non-stochastic input nodes satisfy a directed global Markov property whose graphical id-separation criterion returns a factorization of Markov kernels, on arbitrary input spaces.

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。

↑